Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
For some people walking around a supermarket is the best exercise they get.
Its also quicker and cheaper than shopping online.
Quicker and cheaper?
I've been online shopping for groceries for years. Its far quicker and cheaper. Can do the order at home on the sofa while watching a TV show and when the groceries arrives its quicker to bring inside than even unloading the car would have been.
As for cost its cheaper I've found. We paid until this pandemic a delivery pass subscription for Asda that cost about 50p a week - and the online prices and offers match the in store ones. For 50p per week I don't need to drive to the supermarket, don't need to spend time shopping and most importantly don't need to drag children around the supermarket saying "I want this, I want that" so we don't end up with toys or sweets that weren't on the shopping list in the basket.
Far cheaper and far more convenient getting it delivered for me at least. We've cancelled our delivery pass for now though as its more important that those who need delivery slots more can get them.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
For anyone interested in the debate over the virus' origins and the possibility of it resulting from virus hunting activities, this is an excellent article from a friend and colleague:
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
Substitutions is the big annoyance with how supermarkets do home delivery. When I order butter, no margarine isn't acceptable. When I order mixed nuts, I don't want ground almonds.
This is the thing that will kill the Supermarkets compared to wholesalers and subscriptions.
It is literally unbelievable that even with a 2 week lead time on my order supermarkets are out of stock of what I ordered.
A rational person would ask how that is at all possible and how any customer could find that acceptable.
And the answer is supermarkets operate and hideously complex logistics chain which is simply incapable of holding stock for a future order.
Ocado manage to do it as they don't have pesky customers walking round the store randomly picking things up.
Most of Morrisons in the UK use the Ocado system and hence substitutions are rare. Up here they pick from the other local supermarket (my town has 2 Morrisons but only a tiny Tesco Local) which means substitutions are unavoidable.
The entire British economy has been shut down, plus god knows how many deferred operations, cancer scans, GP checks etc etc. for a model that the university will not release.
"A request for the original code has been made 8 days ago but ignored, it will probably take some kind of legal compulsion to make them release it. Clearly Imperial are too embarrassed by the state of it to ever release it of their own free will, which is unacceptable given it was paid for by the taxpayer and belongs to them."
I have friend and family of all political persuasions who are ranting about the bonking boffin
I kinda pity him. The only way he can now win is if Sweden suddenly dies en masse
It seems a "derivative" of the code has been released. The takedown on it is staggering in that article.
If this stands up, then this is massive. Front page massive.
e.g "I’ll illustrate with a few bugs. In issue 116 a UK “red team” at Edinburgh University reports that they tried to use a mode that stores data tables in a more efficient format for faster loading, and discovered to their surprise that the resulting predictions varied by around 80,000 deaths after 80 days"
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Most parents with school age children would surely be low risk?
Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
For some people walking around a supermarket is the best exercise they get.
Its also quicker and cheaper than shopping online.
Quicker and cheaper?
I've been online shopping for groceries for years. Its far quicker and cheaper. Can do the order at home on the sofa while watching a TV show and when the groceries arrives its quicker to bring inside than even unloading the car would have been.
As for cost its cheaper I've found. We paid until this pandemic a delivery pass subscription for Asda that cost about 50p a week - and the online prices and offers match the in store ones. For 50p per week I don't need to drive to the supermarket, don't need to spend time shopping and most importantly don't need to drag children around the supermarket saying "I want this, I want that" so we don't end up with toys or sweets that weren't on the shopping list in the basket.
Far cheaper and far more convenient getting it delivered for me at least. We've cancelled our delivery pass for now though as its more important that those who need delivery slots more can get them.
Each to their own.
Yeah you're mileage may vary - certainly for us its cheaper without pester power from the children, but others wouldn't have that factor to contend with.
Though I genuinely can't understand how driving to the supermarket, walking around it getting everything from the shelves, driving back and unloading the vehicle can be quicker than browsing a website from the comfort of your own home, answering the doorbell and bringing it in.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
I have code that is 15 years old and just works - I would rewrite things if I needed to but why would you.
Because in terms modelling / ML, the available toolset now available is just so much better. As the link says, no multi-thread, no CUDA, this stuff is just basic core to writing complex models these days. Before you even get into machine learning libraries like using Tensorflow or Pytorch, which not only do neural networks, but contain well written and state of the art optimisers.
Also, I dont know about this specific field, but i would presume ideas have also evolved.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
Specifically with regard to working from home, after all the excited noises coming from the Government's climate change boffins in recent days, you would've thought that this would be all the more reason for ministers to encourage WFH to continue forever, where practical.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
Interest free transition loans similar to how student loans used to be. Support people to retrain in a new area including their living expenses to study full time etc. Initially focus on sectors that aren't going to bounce back.
Actual cost -> probably not that much given current borrowing rates. Repayments taken as an extra X% on salaries over 20k...
Helps us build the workforce we need for a post virus world.
A lot probably won't ever be paid back, but it's cheaper than furloughing people in jobs that will never come back.
Could even be a vehicle for social mobility. Might even call it a national education service?
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
I have code that is 15 years old and just works - I would rewrite things if I needed to but why would you.
Because in terms modelling / ML, the available toolset now available is just so much better. As the link says, no multi-thread, no CUDA, this stuff is just basic core to writing complex models these days. Before you even get into machine learning libraries like using Tensorflow or Pytorch, which not only do neural networks, but contain well written and state of the art optimisers.
Also, I dont know about this specific field, but i would presume ideas have also evolved.
Prior to January this year when was the last time the model was used?
Now I'm not arguing the point here but even though I work in IT I don't know many people who will willingly rewrite something unless there is a real need to do so.
Especially if time is not a priority when running the model and you don't wish to accidently change it.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Most parents with school age children would surely be low risk?
I've always taken my daughter to school as my working hours before this weren't 9-5 so I could but I'd say ~25% of the people I saw at the school gates would be grandparents rather than parents.
Did no one on SAGE ask for the model code to be analysed?
I want to know did anybody challenge the idea for the NHS app to not go with Google / Apple API, i dont know anybody with a technical background who thinks it was a good idea.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
I mean I've no idea but weren't schools the very last thing to be closed down? They were happily playing in the playground while @eadric was crapping himself under the kitchen table. How many CV19 deaths of children have there been?
From the fateful herd immunity speech onwards attendance numbers in schools cratered, by the end half, less than half or so. Before that very nervy and concerned, pretty much everything cancelled outside of class time, a bare bones operation.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
I have code that is 15 years old and just works - I would rewrite things if I needed to but why would you.
Because in terms modelling / ML, the available toolset now available is just so much better. As the link says, no multi-thread, no CUDA, this stuff is just basic core to writing complex models these days. Before you even get into machine learning libraries like using Tensorflow or Pytorch, which not only do neural networks, but contain well written and state of the art optimisers.
Also, I dont know about this specific field, but i would presume ideas have also evolved.
Prior to January this year when was the last time the model was used.
But, as I said previously, this is this guys area of world expertise. Thus, you would hope that would mean he was constantly working and developing the ideas and thus his pandemic model.
Also, as stated previously, it is fairly common practice in academia, that if you have a set of tools that are seen to be really important, that the very least academics like to do is get some minions, in the form of a post-doc or summer intern, to modernize / improve code.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
You appear to have missed out a lot of other parts to the equation.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Specifically with regard to working from home, after all the excited noises coming from the Government's climate change boffins in recent days, you would've thought that this would be all the more reason for ministers to encourage WFH to continue forever, where practical.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
I have to say my bosses are looking long and hard at our office space and the question is being asked as to whether we need it. If we can hire meeting rooms on an hourly or daily basis that will be fine as long as they have all the video conferencing infrastructure.
The word from my team meeting is we may never go back to the office. It's something I'm hearing from friends. Whether we like it or not, working from home or as @NerysHughes called it earlier "working at home", is here to stay.
It may also mean a less formal working atmosphere - I don't go as far as those who work in their Rupert Bear jimjams but the days of a suit and tie seem remote and the justification for them equally pointless.
This is the thing that will kill the Supermarkets compared to wholesalers and subscriptions.
It is literally unbelievable that even with a 2 week lead time on my order supermarkets are out of stock of what I ordered.
A rational person would ask how that is at all possible and how any customer could find that acceptable.
And the answer is supermarkets operate and hideously complex logistics chain which is simply incapable of holding stock for a future order.
Its true the supermarkets are complex - its the reason why their operating costs are £fucktons compared to simpler business models like Aldi. And some of the supermarkets are more inefficient than others. However, its not as simple as you suggest: 1. Sales are pogoing, some of the week on week uplifts were insane and then flipped into high double digit declines and back again. I could have made any volume forecast you like and had data to back it up. Not just that but *what* people were buying swung massively. Combine that with a collapse in ingredient and packaging supplies and you have a problem... 2. They have just done a BIG expansion in online capacity. Morrisons as one example are now using all kind of new stores as pick locations for online, with newly hired and hastily trained staff working at pace. 3. You order something for delivery in 2 weeks - that is NOT a demand order to the manufacturer. You just booked a pick slot for 2 weeks time, with the order to the in store team only generated on that day. Why might the product you want be out of stock? See points 1 and 2 4. Wholesalers who supply foodservice as supplying products intended for restaurants, employee feeding, school feeding etc etc. A lot of these products aren't the cuts / pack formats consumers want, so if you're lucky and have found a wholesaler who has what you want, fabulous. 5. If everyone shops online the supermarkets are finished. A huge cost to serve that currently costs them money every delivery. Infrequent shops with minimal opportunity to uptrade you with offers for things you don't have on your list which is how they have been so successful over the last decade.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
I have code that is 15 years old and just works - I would rewrite things if I needed to but why would you.
Because in terms modelling / ML, the available toolset now available is just so much better. As the link says, no multi-thread, no CUDA, this stuff is just basic core to writing complex models these days. Before you even get into machine learning libraries like using Tensorflow or Pytorch, which not only do neural networks, but contain well written and state of the art optimisers.
Also, I dont know about this specific field, but i would presume ideas have also evolved.
Prior to January this year when was the last time the model was used?
Now I'm not arguing the point here but even though I work in IT I don't know many people who will willingly rewrite something unless there is a real need to do so.
Especially if time is not a priority when running the model and you don't wish to accidently change it.
That isn't how academia is supposed or in my experience works. If your area of expertise is pandemic modelling, you would be expected to be constantly developing, well, a pandemic model.
The world of research doesn't stand still, and when it comes to mathematical modelling and machine learning, the leap and bounds in the tools (by this I mean advances in algorithms and the software) available over the past 5+ years are enormous.
This is quite different to IT industry, where you take time and energy developing specs for a project, such that when it is done it should continue to work as is for a long time to come.
The entire British economy has been shut down, plus god knows how many deferred operations, cancer scans, GP checks etc etc. for a model that the university will not release.
"A request for the original code has been made 8 days ago but ignored, it will probably take some kind of legal compulsion to make them release it. Clearly Imperial are too embarrassed by the state of it to ever release it of their own free will, which is unacceptable given it was paid for by the taxpayer and belongs to them."
This, predicted 40000 deaths for Sweden claim about the model.
I've only seen it in the last day or so. What is the parameters for this.
Is this some bullshit "no lockdown" setup?
I would guess so,
I've found the paper and it's not. It has modeled as best it can the Swedish approach but there are several important things missing. Bans in gatherings of over 50 people and ban on flights from outside the EEA.
Also the Imperial Model has assumption built in that Sweden breaks. Almost 40% of Sweden's homes are occupied by a single person.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
academics make truly sh!t developers. they really do. just the fact it is a 15K line undocumented block of C.... you don't need to read any further. it is broken, wrong, buggy, wrong (again), and utterly utterly shit.
Whatever your profession, lawyer, journalist, shop keeper. Imagine you just read that someone purported to be an expert and proceeded to make *every* *single* one of the rookie 'we-all-make-this-fuck-when-we-first-start-ha-ha' mistakes. And then crashed the world. You'd be a bit pissed off. Which I am. I hope everyone who touched this shit show never works again.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Most parents with school age children would surely be low risk?
I am a very old parent of young children (57 years older than my youngest child). So not that low risk. But I am not very worried about catching it via my children. I am much more worried about catching it off other parents or teachers taking and collection in the narrow corridor where we have to wait. I would like my children to go back to nursery next week, they are suffering. But it doesn't have to be 5 days, why not split it into two, one half Mon-Wed, the other half Thurs-Sat. And stagger start and collection time please.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
academics make truly sh!t developers. they really do. just the fact it is a 15K line undocumented block of C.... you don't need to read any further. it is broken, wrong, buggy, wrong (again), and utterly utterly shit.
Whatever your profession, lawyer, journalist, shop keeper. Imagine you just read that someone purported to be an expert and proceeded to make *every* *single* one of the rookie 'we-all-make-this-fuck-when-we-first-start-ha-ha' mistakes. And then crashed the world. You'd be a bit pissed off. Which I am. I hope everyone who touched this shit show never works again.
I don't disagree, especially self taught scientists from non-computer science fields. I wrote horrible code when I did my PhD, not 15k lines of undocumented C, but it wasn't good. But then I wasn't writing a model to simulate a global pandemic.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
Is it their "patriotic duty" or would you support the idea of parents who keep their children at home being prosecuted and fined?
Only if they are failing to home school their children effectively
About as shocking as finding out that Ed Miliband was only allowed to play the marxist version of monopoly as a kid
This bit made me think our app will be as useful as a condom with holes in both ends.
One person familiar with the NHS testing process said that its app was able to work in the background in most cases, except when two iPhones were locked and left unused for around 30 minutes, without any Android devices coming within 60m of the devices. Bringing an Android device running the NHS app close to the iPhone would “wake up” its Bluetooth connection, this person said.
Switching to Google and Apple’s contact technology would avoid such issues but likely mean the UK had to abandon its centralised database, which representatives of the two tech companies have indicated is incompatible with their “decentralised” approach.
What is the point of the app? It just sounds like someone in government heard that apps are cool so determined we should have one, without much consideration of what it was supposed to do, let alone how it should do it.
Hancock loves Apps.
For the next phase we do need Test and Trace and Isolate.
Why not subcontract the whole shebang to South Korea. I don't think those in Seoul are bothered by where I am, nor want to use me in their big data.
Indeed with a Samsung phone they probably know already...
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
I can only presume he was being very generous.
Pulling a stunt of sticking all your massive project in one mega file will get you a very bad mark even at first year undergrad computer science classes. Before you even get to no documentation.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
+1 - if John Carmack is saying the code is fine you need to be a brave man to claim it isn't.
It's why I'm not concerned about that nor the single 15,000 line file a single 15,000 line class would be an issue, a correctly written program in a single file isn't perfect but it's just a bad decision rather than an issue.
Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
Substitutions is the big annoyance with how supermarkets do home delivery. When I order butter, no margarine isn't acceptable. When I order mixed nuts, I don't want ground almonds.
This is the thing that will kill the Supermarkets compared to wholesalers and subscriptions.
It is literally unbelievable that even with a 2 week lead time on my order supermarkets are out of stock of what I ordered.
A rational person would ask how that is at all possible and how any customer could find that acceptable.
And the answer is supermarkets operate and hideously complex logistics chain which is simply incapable of holding stock for a future order.
I totally agree.
We use supermarkets as a top up, maybe once a month, for tinned goods etc. Our food gets delivered by Abel and Cole. Fresh and organic meat, fish, veg, fruit, cheese. Yeh we pay a bit more, but they've never missed a beat during this crisis.
My parents, who have been supermarket stalwarts my whole life, are astounded with the difference switching to their local farm shop which now delivers.
Good food small businesses, farm shops, delis, butchers, bakers etc are going to have a boon after all this has shaken out.
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
--AS
There's a lot of blame because well, um, well Experts...
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
My question about Ferguson model. This is thus guys area of expertise, hiw the hell do you go 13 years never improving it / modernizing it. In my time in academia, if nothing else that is what you found some grant money to get post-grads / post-docs to do.
Also, it isnt like he hasnt had use for it. He has been called to employ it several tines over the years.
academics make truly sh!t developers. they really do. just the fact it is a 15K line undocumented block of C.... you don't need to read any further. it is broken, wrong, buggy, wrong (again), and utterly utterly shit.
Whatever your profession, lawyer, journalist, shop keeper. Imagine you just read that someone purported to be an expert and proceeded to make *every* *single* one of the rookie 'we-all-make-this-fuck-when-we-first-start-ha-ha' mistakes. And then crashed the world. You'd be a bit pissed off. Which I am. I hope everyone who touched this shit show never works again.
I don't think his models/predictions have been that different from others in terms of the shutdown. If anything, it seems as though it didn't go far enough if the effect of flights was from him.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
I can only presume he was being very generous.
Pulling a stunt of sticking all your massive project in one mega file will get you a very bad mark even at first year undergrad computer science classes. Before you even get to no documentation.
This is stuff is literally programming 101.
The world runs on codebases like this. Most software, even in quite critical systems, is crap; and a lot of it will not have been touched in many years, maybe even decades.
Of course you would hope software was better written, but it's the model that matters most, not the implementation, even if it it quite poor.
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
--AS
A very good point. If the predictions were totally out of whack, rival groups would have made a lot more noise by now.
I am not an epidemiologist but surely in any model its not the code that matters so much its the assumptions?
You can have an amazing model but if your assumptions are flawed it will be just a case of garbage in, garbage out.
I worry when people defer too much to a model, any model, without thinking about the assumptions that underpin it.
Similarly the criticism should be centered on the substance (e.g. formulae/relations used), not style (e.g. parallelised or not)
No, criticism of not using parallelised, using modern optimization, etc, is valid. How long does his model take to run? I am sure it would be orders of magnitude faster if he was using modern techniques. And also, with many modern optimization techniques, you would be able to "search" and "test" a wider range of parameters.
Specifically with regard to working from home, after all the excited noises coming from the Government's climate change boffins in recent days, you would've thought that this would be all the more reason for ministers to encourage WFH to continue forever, where practical.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
I have to say my bosses are looking long and hard at our office space and the question is being asked as to whether we need it. If we can hire meeting rooms on an hourly or daily basis that will be fine as long as they have all the video conferencing infrastructure.
The word from my team meeting is we may never go back to the office. It's something I'm hearing from friends. Whether we like it or not, working from home or as @NerysHughes called it earlier "working at home", is here to stay.
It may also mean a less formal working atmosphere - I don't go as far as those who work in their Rupert Bear jimjams but the days of a suit and tie seem remote and the justification for them equally pointless.
Exactly. It won't work for everyone, but for an awful lot of people there's probably no point in going back to the way things were. Quite apart from the cost savings to businesses, the benefits to both employees and the environment from doing away with all that commuting that will no longer have to happen will be enormous. An entire world of opportunities that didn't exist before has should open up to people with mobility issues, now that they're relieved of the need to travel to access them.
Moreover, for those who do still have to go out to work, getting a large fraction of the traffic off the roads and the bulk of the commuters off the trains will make life so much less stressful, and save millions of hours (and an appreciable number of lives) every year from all those traffic jams, road accidents, broken down trains and "train cancelled due to no driver available" episodes that will no longer happen.
I frankly don't see who loses from all of this except the public transport providers. Especially in the current climate, very few voters will be troubled if they all end up being quietly renationalised and run under something like the TfL model. It's a no brainer.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
I can only presume he was being very generous.
Pulling a stunt of sticking all your massive project in one mega file will get you a very bad mark even at first year undergrad computer science classes. Before you even get to no documentation.
This is stuff is literally programming 101.
The world runs on codebases like this. Most software, even in quite critical systems, is crap; and a lot of it will not have been touched in many years, maybe even decades.
Of course you would hope software was better written, but it's the model that matters most, not the implementation, even if it it quite poor.
I disagree. The point of academia is constant progression and improvement. By not using modern tools available (and by this I mean algorithms and actual software libraries), you are missing out on an enormous opportunities for potential improvements.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
Mathematically modelling an infectious disease epidemic is actually fairly simple.
The errors are not likely to be in the maths though, but rather the assumptions. What is the R number unrestrained? What percentage of people will comply with lockdown? What is the prevalence of the disease in the community? There are too many things that we simply do not know yet.
The entire British economy has been shut down, plus god knows how many deferred operations, cancer scans, GP checks etc etc. for a model that the university will not release.
"A request for the original code has been made 8 days ago but ignored, it will probably take some kind of legal compulsion to make them release it. Clearly Imperial are too embarrassed by the state of it to ever release it of their own free will, which is unacceptable given it was paid for by the taxpayer and belongs to them."
I am not an epidemiologist but surely in any model its not the code that matters so much its the assumptions?
You can have an amazing model but if your assumptions are flawed it will be just a case of garbage in, garbage out.
I worry when people defer too much to a model, any model, without thinking about the assumptions that underpin it.
Similarly the criticism should be centered on the substance (e.g. formulae/relations used), not style (e.g. parallelised or not)
No, criticism of not using parallelised, using modern optimization, etc, is valid. How long does his model take to run? I am sure it would be orders of magnitude faster if he was using modern techniques. And also, with many modern optimization techniques, you would be able to "search" and "test" a wider range of parameters.
If they needed to parallelise it, they would have. I have tons of code that works fine on a single thread, even if it takes an hour or so. And remember that not every thing can be easily parallelised, I think the only thing might be running multiple models with different initial conditions. You don't need to parellelise the code for that, you just run it a bunch of times in the shell.
Edit: when I said style shouldn't be critisied, I mean in a way that seeks to suggest that the model isn't accurate just because it doesn't use parellelisation or whatever.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
+1 - if John Carmack is saying the code is fine you need to be a brave man to claim it isn't.
It's why I'm not concerned about that nor the single 15,000 line file a single 15,000 line class would be an issue, a correctly written program in a single file isn't perfect but it's just a bad decision rather than an issue.
If you want your Big Mac fix and don't mind queuing half the night for the privilege so be it but will parents be more interested in the health of their children (and themselves) than in doing their patriotic duty? It may well be given how subservient we were going into lockdown, Johnson and Cummings think we will be equally subservient coming out of it - we'll see.
What was the Parent Teachers Association group for the UK (now called Parentkind) had a survey out to all PTA’s and it had been promoted by the DfE. Results published fully tomorrow but, as per the BBC news, out of a quarter of a million responses 90% of parents would not be happy sending their children to school after lockdown ends. I’ve kept saying this and maybe it might get through now, government can’t afford to announce something that will be so comprehensively ignored. The lack of trust that stems from the late lockdown and the attempt at vaccine free herd immunity is palpable.
So what do those parents want to do instead ?
If your children are at home you can't go to work, hence your furlough / wfh may continue.
It could also be that some parents are fond of their children and do not want them to die from exotic bat viruses; unlikely after being locked down wiv da kids for three weeks but not impossible.
The death rate for under 18s is virtually 0 from Covid, unless kids live with grandparents they should be going back to school as German kids are starting to do
The issue with the schools is not the kids, its the parents (and grandparents).
Most parents with school age children would surely be low risk?
I am a very old parent of young children (57 years older than my youngest child). So not that low risk. But I am not very worried about catching it via my children. I am much more worried about catching it off other parents or teachers taking and collection in the narrow corridor where we have to wait. I would like my children to go back to nursery next week, they are suffering. But it doesn't have to be 5 days, why not split it into two, one half Mon-Wed, the other half Thurs-Sat. And stagger start and collection time please.
The original code was “a single 15,000 line file that had been worked on for a decade”
WTF....
Despite that John Carmack, who worked on the refactoring, was quite complemenatary about the code. If Carmack doesn't have any major complaints about the code — which is an entirely different issue from the modelling — that is more than good enough for me.
+1 - if John Carmack is saying the code is fine you need to be a brave man to claim it isn't.
It's why I'm not concerned about that nor the single 15,000 line file a single 15,000 line class would be an issue, a correctly written program in a single file isn't perfect but it's just a bad decision rather than an issue.
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
--AS
A very good point. If the predictions were totally out of whack, rival groups would have made a lot more noise by now.
I am not an epidemiologist but surely in any model its not the code that matters so much its the assumptions?
You can have an amazing model but if your assumptions are flawed it will be just a case of garbage in, garbage out.
I worry when people defer too much to a model, any model, without thinking about the assumptions that underpin it.
Similarly the criticism should be centered on the substance (e.g. formulae/relations used), not style (e.g. parallelised or not)
No, criticism of not using parallelised, using modern optimization, etc, is valid. How long does his model take to run? I am sure it would be orders of magnitude faster if he was using modern techniques. And also, with many modern optimization techniques, you would be able to "search" and "test" a wider range of parameters.
If they needed to parallelise it, they would have. I have tons of code that works fine on a single thread, even if it takes an hour or so. And remember that not every thing can be easily parallelised, I think the only thing might be running multiple models with different initial conditions. You don't need to parellelise the code for that, you just run it a bunch of times in the shell.
And as I say, it isn't just that, there are many modern optimization algorithms for example calculating hyperparameters of models. They weren't about 13 years ago, but allow one to investigate the model space far more efficiently.
Not going to get into the real weeds with the technicalities of machine learning and mathematical modelling, other than to say it is a very different world from 2007.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
We'd do that NOW, in fact they have done to an extent, but do you really think anyone, say last October, was thinking "we really need to urgently review and refactor our pandemic modelling software"? I don't, it wasn't considered valuable to any degree back then.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
We'd do that NOW, in fact they have done to an extent, but do you really think anyone, say last October, was thinking "we really need to urgently review and refactor our pandemic modelling software"? I don't, it wasn't considered valuable to any degree back then.
As I said previously, his specialist field is modelling pandemics. This isn't how academia is supposed to work, you should be constantly evolving and developing your work...which by all accounts Ferguson is the man who the government go to come an outbreak, and done so several times over the past 20 years.
It isn't just refactoring your code, it is taking advantage of a modern tools and I would bloody hope that in his field algorithms have also evolved.
Not sure how to do a DM, but thanks very much to @eek for his tips on the last thread. I've checked whether I had Norton installed (really not, though I used to have Norton virus protection), then reset the browser settings as you suggested, then had a look at Paypal and lo and behold, the warning message has vanished.
Specifically with regard to working from home, after all the excited noises coming from the Government's climate change boffins in recent days, you would've thought that this would be all the more reason for ministers to encourage WFH to continue forever, where practical.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
I have to say my bosses are looking long and hard at our office space and the question is being asked as to whether we need it. If we can hire meeting rooms on an hourly or daily basis that will be fine as long as they have all the video conferencing infrastructure.
The word from my team meeting is we may never go back to the office. It's something I'm hearing from friends. Whether we like it or not, working from home or as @NerysHughes called it earlier "working at home", is here to stay.
It may also mean a less formal working atmosphere - I don't go as far as those who work in their Rupert Bear jimjams but the days of a suit and tie seem remote and the justification for them equally pointless.
Exactly. It won't work for everyone, but for an awful lot of people there's probably no point in going back to the way things were. Quite apart from the cost savings to businesses, the benefits to both employees and the environment from doing away with all that commuting that will no longer have to happen will be enormous. An entire world of opportunities that didn't exist before has should open up to people with mobility issues, now that they're relieved of the need to travel to access them.
Moreover, for those who do still have to go out to work, getting a large fraction of the traffic off the roads and the bulk of the commuters off the trains will make life so much less stressful, and save millions of hours (and an appreciable number of lives) every year from all those traffic jams, road accidents, broken down trains and "train cancelled due to no driver available" episodes that will no longer happen.
I frankly don't see who loses from all of this except the public transport providers. Especially in the current climate, very few voters will be troubled if they all end up being quietly renationalised and run under something like the TfL model. It's a no brainer.
I’m expecting a major uptick in home schooling, either for those students with vulnerable health or their parents, probably wider than that. The need for specialist private online tuition is a possible opening to complement that. Maybe a local hub for socialisation, bringing students together. There are lot of opportunities here.
The main criticism of Neil Ferguson's model seems to be that it does not produce replicable results.
non replicable results are quite common in complex system modelling. You might have a rank function buried in a branch and the rank hits a tie. Each time you run it you may go down different branches with different results. So you have MCMC approaches which basically mean you run it gazillions of times and take an average. But the issues here are deeper.
If different hardware is producing different results, yeah, you are in a world of pain.
To the earlier point about garbage in garbage out GIGO. Absolutely right and the assumptions and parameters are critical, but you can also get code that itself creates garbage as well as algorithmic approaches that do the same. Models are an exquisite balance of data, assumptions, algorithms, implementations, interpretations and human judgement. A person decides all this. Someone decides the framework, someone picks the data variables, granulairity and derivations. Someone decides what to do with outliers, with nulls, with NaN. Someone creates the decision space, someone picks the language. This is all still a deeply human process.
Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
Substitutions is the big annoyance with how supermarkets do home delivery. When I order butter, no margarine isn't acceptable. When I order mixed nuts, I don't want ground almonds.
This is the thing that will kill the Supermarkets compared to wholesalers and subscriptions.
It is literally unbelievable that even with a 2 week lead time on my order supermarkets are out of stock of what I ordered.
A rational person would ask how that is at all possible and how any customer could find that acceptable.
And the answer is supermarkets operate and hideously complex logistics chain which is simply incapable of holding stock for a future order.
I totally agree.
We use supermarkets as a top up, maybe once a month, for tinned goods etc. Our food gets delivered by Abel and Cole. Fresh and organic meat, fish, veg, fruit, cheese. Yeh we pay a bit more, but they've never missed a beat during this crisis.
My parents, who have been supermarket stalwarts my whole life, are astounded with the difference switching to their local farm shop which now delivers.
Good food small businesses, farm shops, delis, butchers, bakers etc are going to have a boon after all this has shaken out.
Unfortunately much of the country doesn't have access to outlets like this, even in relatively well-off areas. All we've got in town is the bakery, which I might patronise but very seldom do because it's only open when I'm stuck at work.
And yes, I'm sure that the doorstep deliveries of fresh food are lovely, but you would quickly find that these firms were far more hopeless than the supermarkets if everyone rushed to use them. Which, of course, they will not because the vast numbers of people living off the basic state pension, universal credit, zero hours contracts and minimum wage crap jobs cannot aspire to such luxuries.
Niche retailers have their place but they'll remain niche. People who have to work don't have the time to spend wandering around sundry specialist retailers in the manner of a pre-modern housewife doing her daily shopping trip to town. People who aren't in work mostly can't afford to shop in these outlets. And most consumers appreciate the convenience and relatively low prices that a supermarket can offer on a wide variety of goods, all brought together in one place.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
We'd do that NOW, in fact they have done to an extent, but do you really think anyone, say last October, was thinking "we really need to urgently review and refactor our pandemic modelling software"? I don't, it wasn't considered valuable to any degree back then.
Why not?
This has been about the 8th epidemic so far this century.
Before rushing to judgement on Ferguson's code, please remember that the LSHTM group, who are also leading epidemiologists, found very similar results using a completely independent implementation and a related but different model, from similar-ish assumptions. Thus it seems very unlikely that some showstopper bug in the Imperial code lead to the conclusions being seriously wrong.
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
--AS
Adam Kucharski (at the London School) claimed in a tweet a while back that the media hype about government turning policy around rapidly in response to the famous Imperial Report 9 ("Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand") was overblown, as LSHTM had been feeding very similar input in to the government as a result of their modelling at around the same time.
Of course, if they were making similar modelling assumptions then the fact the two largely agreed doesn't tell you all that much about the correctness of their conclusions, though it does reduce the probability this is all just down to a bug in one man's code. I've seen plenty of shoddy analytics code in corporate environments for what it's worth but it's often ugly, undocumented, inefficient and despite all that, pretty much works, as accurately as it needed to. Would be rather more reassuring if lockdown was based on transparent, documented, well-written code of course. I think the LSHTM group have been good at sticking their code online throughout the pandemic response (here's their repository) though to far less fanfare than Imperial managed. Which reminds me of something I posted last night...
Being Private Eye everything has to be taken with a certain modicum of salt but even so, their foot and mouth special report is an interesting read. Very telling how the Imperial modelling team at the time (presumably largely or entirely via Roy Anderson, so this isn't necessarily a direct comment on Ferguson himself who was the junior partner) carefully cultivated the media not just to improve the research group's image/standing but to try to "bounce" policy decisions when rivals were suggesting the evidence favoured alternative approaches.
(The eagle-eyed may note Christl Donnelly being quoted a lot in pieces about her colleague Ferguson at present! And recall the Oxford epidemiologist Sunetra Gupta for her criticism of the Imperial team's modelling assumptions earlier in the epidemic, a bit - or rather, a heck of a lot - of history there...)
Doesn't detract from the fact Anderson / Ferguson are high-profile figures who have made very important contributions to the field. But ability to play the media is a handy skill to have if you're involved in policy-determining research.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
We'd do that NOW, in fact they have done to an extent, but do you really think anyone, say last October, was thinking "we really need to urgently review and refactor our pandemic modelling software"? I don't, it wasn't considered valuable to any degree back then.
Why not?
This has been about the 8th epidemic so far this century.
Indeed, since SARS and MERS, the entire world of health security has been on high alert to another novel coronavirus with pandemic potential.
The main criticism of Neil Ferguson's model seems to be that it does not produce replicable results.
For me, as I understand it is the modal amongst other things predicted that the hospitalises would be buissy to the point of overflowing, there for the dicition was made to puss people out the door ASAP. which while ok for some, meant that a lot of people were returned to care homes bringing the virus with them. and this has resulted in a large part of the deaths.
The lockdown has already devastated cashflows. Even with the furlough scheme, businesses have many fixed costs and far less (if any) income coming in.
Yet if businesses do open up, demand is likely to be low. Partly because people are afraid to go out, partly because they are afraid to spend money.
That means that when we do re-open, it is incredibly unlikely 100% of those furloughed staff will be needed. Perhaps 70% at best. And I'm guessing a whole lot less, at least for the next few months.
In one of my whatsapp groups, my friends laughed at me because I'm still working when they are all enjoying a free holiday on £2500 a month.
So I said, at least it proves my employer knows I'm absolutely necessary to the business. How many of you can say the same? How many of you think you will have jobs to go back to when this ends?
That shut them up.
And there is the problem. When we restart the economy, it will be slow. It may take a year or more to return to previous levels of demand.
Furlough has lulled most people into a false sense of security. Once the rug is pulled out from underneath them, that's when the fun begins.
Oh yeah. Middle class supposedly comfortable people trying to claim UC like all those skivers and scroungers they read about in the Daily Mail. Finding that it isn't the dossers cakewalk made out to be...
Yep, exactly.
Most of the people who are about to find out UC is £74 a week probably never thought they were the sort of person who'd be claiming benefits.
They had all the education, all the skills - transferable skills too - and of course all the contacts. They never expected to be out of work for more than a couple of weeks in their life, if they ever expected it at all.
A lot of people are about to find out how the other half lives very soon. And it won't be pretty. Or it will be delicious. Depending on your point of view.
A lot of people won't qualify for the £74 a week as their partner will be earning just enough for them not to qualify.
The government upped the single allowance to £94 per week for the next year, presumably so those having to claim won't realise quite how low it has been. £137 p.w. for a couple.
Every £1 of income (e.g. from one partner working) reduces that by 63p.
@FrancisUrquhart, I greatly value your contributions. But I think you have too optimistic a view of academic code.
I still run old code when it implements something I need, even if it could be refined with modern techniques. It's quicker to do that than refactor and rebugfix. If it needs to run faster, throw money at AWS. Indeed, when it's battle-tested code, I prefer not to fiddle with it. My experience is that each grant is different, and tends to need new code doing something novel rather than a refinement of old code. And unless it's funded, there is neither time nor money to refactor (I rarely touch code myself any more, leaving that for postdocs and students!).
If the government asked my advice (on something within my expertise), I'd most likely run old code. Maybe even 13 year-old code. I myself would be uneasy if my results were then the only thing leading to policy, but if my peers were coming to similar conclusions then I'd be pretty confident that we collectively got the right answers from the inputs we had.
Specifically with regard to working from home, after all the excited noises coming from the Government's climate change boffins in recent days, you would've thought that this would be all the more reason for ministers to encourage WFH to continue forever, where practical.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
I have to say my bosses are looking long and hard at our office space and the question is being asked as to whether we need it. If we can hire meeting rooms on an hourly or daily basis that will be fine as long as they have all the video conferencing infrastructure.
The word from my team meeting is we may never go back to the office. It's something I'm hearing from friends. Whether we like it or not, working from home or as @NerysHughes called it earlier "working at home", is here to stay.
It may also mean a less formal working atmosphere - I don't go as far as those who work in their Rupert Bear jimjams but the days of a suit and tie seem remote and the justification for them equally pointless.
I've been WFH for years now and it is OK but some firms may be surprised that there are swings as well as roundabouts. No water-cooler moments, shared tribal knowledge, team cohesion etc.
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
We'd do that NOW, in fact they have done to an extent, but do you really think anyone, say last October, was thinking "we really need to urgently review and refactor our pandemic modelling software"? I don't, it wasn't considered valuable to any degree back then.
Why not?
This has been about the 8th epidemic so far this century.
Indeed, since SARS and MERS, the entire world of health security has been on high alert to another novel coronavirus with pandemic potential.
But it seems not on high enough alert to have a good poke around in the modelling software we were relying on.
I am a very old parent of young children (57 years older than my youngest child). So not that low risk. But I am not very worried about catching it via my children. I am much more worried about catching it off other parents or teachers taking and collection in the narrow corridor where we have to wait. I would like my children to go back to nursery next week, they are suffering. But it doesn't have to be 5 days, why not split it into two, one half Mon-Wed, the other half Thurs-Sat. And stagger start and collection time please.
Thanks v interesting.
There's quite a bit of scope for looking at a part-time economy as a step back. At a micro level, I'm the exec portfolio-holder for car parks in my borough. We suspended charges dyring the virus peak (nobody except carers and nurses were using them and it seemed inappropriate to charge them) but when people are driving again we'll need to reimpose them as they're a crucial part of council income.
But instead of just restoring season tickets (traditionally bought by employees comnmuting in), I'm thinking of offering a voucher scheme, whereby employers could buy books of vouchers or their electronic equivalent to give to staff who came in two or three times a week. So businesses that can't yet afford to have their staff all turn up every day could ve encouraged to have half for two days, the other half for another two, or whatever worked best for them
Hmm If John Carmack reckons the code is OK it's probably OK. Seems it's ineffficient and atrociously documented, possibly a bit buggy - he's definitely an academic then
Nationally though furlough is likely to continue for some time, even at a reduced rate as the Government continues to pursue watered down social democracy
We have been sourcing meat, fruit, veg and other items via wholesalers who would normally deliver to the restaurant trade but have found a lucrative opening in the home delivery market.
Our second meat order arrived today - prompt and looks excellent quality. Mostly but not wholly British as well.
Throw in subscription delivery services and I see the supermarkets losing their high value customers once this is all over.
The wholesalers will be able to keep home deliveries, their customers will love the price/quality over the supermarkets. They'll never have substitution issues as the wholesalers deal in actual stock and don't have the supermarkets fragile JIT structures.
The well heeled middle class never need to bother a supermarket again.
I'd say the elderly didn't previously take up home delivery as shopping filled time in the retirement calendar.
They may not bother to return.
I am 66 and i won't be going into a big supermarket anytime soon Click and collect or home delivery for us from now on until we feel safe ir have been vaccinated...
Seems an utter waste of time attending a supermarket if you can order online at the best of times.
For some people walking around a supermarket is the best exercise they get.
Its also quicker and cheaper than shopping online.
Quicker and cheaper?
I've been online shopping for groceries for years. Its far quicker and cheaper. Can do the order at home on the sofa while watching a TV show and when the groceries arrives its quicker to bring inside than even unloading the car would have been.
As for cost its cheaper I've found. We paid until this pandemic a delivery pass subscription for Asda that cost about 50p a week - and the online prices and offers match the in store ones. For 50p per week I don't need to drive to the supermarket, don't need to spend time shopping and most importantly don't need to drag children around the supermarket saying "I want this, I want that" so we don't end up with toys or sweets that weren't on the shopping list in the basket.
Far cheaper and far more convenient getting it delivered for me at least. We've cancelled our delivery pass for now though as its more important that those who need delivery slots more can get them.
Each to their own.
Yeah you're mileage may vary - certainly for us its cheaper without pester power from the children, but others wouldn't have that factor to contend with.
Though I genuinely can't understand how driving to the supermarket, walking around it getting everything from the shelves, driving back and unloading the vehicle can be quicker than browsing a website from the comfort of your own home, answering the doorbell and bringing it in.
I don't drive to supermarkets, I drive past supermarkets between work and home.
Not sure how to do a DM, but thanks very much to @eek for his tips on the last thread. I've checked whether I had Norton installed (really not, though I used to have Norton virus protection), then reset the browser settings as you suggested, then had a look at Paypal and lo and behold, the warning message has vanished.
PayPal also supports Multi-factor authentication, so somebody would need more than your password to login.
Adam Kucharski (at the London School) claimed in a tweet a while back that the media hype about government turning policy around rapidly in response to the famous Imperial Report 9 ("Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand") was overblown, as LSHTM had been feeding very similar input in to the government as a result of their modelling at around the same time.
Yes, I've been following them. I rather prefer their less flashy approach to the science. Doing it quietly right. I share your implied distaste for the PR-first approach of the Imperial crew. I must admit a little schadenfreude at the recent headlines
(The eagle-eyed may note Christl Donnelly being quoted a lot in pieces about her colleague Ferguson at present! And recall the Oxford epidemiologist Sunetra Gupta for her criticism of the Imperial team's modelling assumptions earlier in the epidemic, a bit - or rather, a heck of a lot - of history there...)
More than commonly known, indeed! Perhaps it will make a fun novel one day.
The main criticism of Neil Ferguson's model seems to be that it does not produce replicable results.
For me, as I understand it is the modal amongst other things predicted that the hospitalises would be buissy to the point of overflowing, there for the dicition was made to puss people out the door ASAP. which while ok for some, meant that a lot of people were returned to care homes bringing the virus with them. and this has resulted in a large part of the deaths.
Increasingly it is the shortage of pharmaceuticals that is becoming problematic. The ICU guys are running short of several staples, and using substitutes. This will be one of the limiting factors alongside PPE related issues in getting urgent surgery done.
The peak was a close run thing, but the limit is going to be skilled personnel, and medicines. These are very sick patients with multiple organ failures. Dialysis stuff is short too.
These are international issues of supply, not particularly of our own governments making.
Adam Kucharski (at the London School) claimed in a tweet a while back that the media hype about government turning policy around rapidly in response to the famous Imperial Report 9 ("Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand") was overblown, as LSHTM had been feeding very similar input in to the government as a result of their modelling at around the same time.
Yes, I've been following them. I rather prefer their less flashy approach to the science. Doing it quietly right. I share your implied distaste for the PR-first approach of the Imperial crew. I must admit a little schadenfreude at the recent headlines
(The eagle-eyed may note Christl Donnelly being quoted a lot in pieces about her colleague Ferguson at present! And recall the Oxford epidemiologist Sunetra Gupta for her criticism of the Imperial team's modelling assumptions earlier in the epidemic, a bit - or rather, a heck of a lot - of history there...)
More than commonly known, indeed! Perhaps it will make a fun novel one day.
--AS
The media narrative has disingenuous by omission from the start. Pretty much every report only refers to Ferguson's model that spat out the disaster scenario, as if nobody had modelled anything before that, hence going for "herd immunity", then White Knight Ferguson turned up and they U-Turned.
When in fact, we know from early January work was being done and those early models used early Chinese data, and it was when more up to date stuff in particular the Italian data, then the 8x capacity horror showed up.
If we are going to see an easing of the lockdown from next week how long do you suppose until the first of the 80+ local council vacancies are filled with by-elections? Which will be first and what will the actual impact of CV19 be on those elections?
Not sure how to do a DM, but thanks very much to @eek for his tips on the last thread. I've checked whether I had Norton installed (really not, though I used to have Norton virus protection), then reset the browser settings as you suggested, then had a look at Paypal and lo and behold, the warning message has vanished.
PayPal also supports Multi-factor authentication, so somebody would need more than your password to login.
I would recommend enabling that feature.
Yes, I think you're right - I hate it usually but with the doubt hovering, I'd better.
Comments
https://thebulletin.org/2020/05/natural-spillover-or-research-lab-leak-why-a-credible-investigation-in-needed-to-determine-the-origin-of-the-coronavirus-pandemic/
Most of Morrisons in the UK use the Ocado system and hence substitutions are rare. Up here they pick from the other local supermarket (my town has 2 Morrisons but only a tiny Tesco Local) which means substitutions are unavoidable.
If this stands up, then this is massive. Front page massive.
e.g "I’ll illustrate with a few bugs. In issue 116 a UK “red team” at Edinburgh University reports that they tried to use a mode that stores data tables in a more efficient format for faster loading, and discovered to their surprise that the resulting predictions varied by around 80,000 deaths after 80 days"
Though I genuinely can't understand how driving to the supermarket, walking around it getting everything from the shelves, driving back and unloading the vehicle can be quicker than browsing a website from the comfort of your own home, answering the doorbell and bringing it in.
Also, I dont know about this specific field, but i would presume ideas have also evolved.
Forcing businesses to adapt using modern technology has revealed that (a) this approach works well or very well in most cases and (b) that they're wasting an awful lot of money on office space that is either completely useless or which they can, at the very least, find workable solutions to coping without. Why such significant improvements to efficiency should be abandoned in 2033 or whenever this shitshow is finally over, God only knows. It would be like people in the 1920s abandoning motorized transport because they felt nostalgic about horses.
I've only seen it in the last day or so. What is the parameters for this.
Is this some bullshit "no lockdown" setup?
Support people to retrain in a new area including their living expenses to study full time etc. Initially focus on sectors that aren't going to bounce back.
Actual cost -> probably not that much given current borrowing rates.
Repayments taken as an extra X% on salaries over 20k...
Helps us build the workforce we need for a post virus world.
A lot probably won't ever be paid back, but it's cheaper than furloughing people in jobs that will never come back.
Could even be a vehicle for social mobility. Might even call it a national education service?
Now I'm not arguing the point here but even though I work in IT I don't know many people who will willingly rewrite something unless there is a real need to do so.
Especially if time is not a priority when running the model and you don't wish to accidently change it.
Also, as stated previously, it is fairly common practice in academia, that if you have a set of tools that are seen to be really important, that the very least academics like to do is get some minions, in the form of a post-doc or summer intern, to modernize / improve code.
We will see later if I am right.
The word from my team meeting is we may never go back to the office. It's something I'm hearing from friends. Whether we like it or not, working from home or as @NerysHughes called it earlier "working at home", is here to stay.
It may also mean a less formal working atmosphere - I don't go as far as those who work in their Rupert Bear jimjams but the days of a suit and tie seem remote and the justification for them equally pointless.
1. Sales are pogoing, some of the week on week uplifts were insane and then flipped into high double digit declines and back again. I could have made any volume forecast you like and had data to back it up. Not just that but *what* people were buying swung massively. Combine that with a collapse in ingredient and packaging supplies and you have a problem...
2. They have just done a BIG expansion in online capacity. Morrisons as one example are now using all kind of new stores as pick locations for online, with newly hired and hastily trained staff working at pace.
3. You order something for delivery in 2 weeks - that is NOT a demand order to the manufacturer. You just booked a pick slot for 2 weeks time, with the order to the in store team only generated on that day. Why might the product you want be out of stock? See points 1 and 2
4. Wholesalers who supply foodservice as supplying products intended for restaurants, employee feeding, school feeding etc etc. A lot of these products aren't the cuts / pack formats consumers want, so if you're lucky and have found a wholesaler who has what you want, fabulous.
5. If everyone shops online the supermarkets are finished. A huge cost to serve that currently costs them money every delivery. Infrequent shops with minimal opportunity to uptrade you with offers for things you don't have on your list which is how they have been so successful over the last decade.
The world of research doesn't stand still, and when it comes to mathematical modelling and machine learning, the leap and bounds in the tools (by this I mean advances in algorithms and the software) available over the past 5+ years are enormous.
This is quite different to IT industry, where you take time and energy developing specs for a project, such that when it is done it should continue to work as is for a long time to come.
Every country with a pandemic has a care home situation it seems.
Also the Imperial Model has assumption built in that Sweden breaks. Almost 40% of Sweden's homes are occupied by a single person.
Whatever your profession, lawyer, journalist, shop keeper. Imagine you just read that someone purported to be an expert and proceeded to make *every* *single* one of the rookie 'we-all-make-this-fuck-when-we-first-start-ha-ha' mistakes. And then crashed the world. You'd be a bit pissed off. Which I am. I hope everyone who touched this shit show never works again.
I would like my children to go back to nursery next week, they are suffering. But it doesn't have to be 5 days, why not split it into two, one half Mon-Wed, the other half Thurs-Sat. And stagger start and collection time please.
https://www.newscientist.com/article/mg15220500-300-league-tables-just-dont-add-up/
If the conclusions are wrong (and we may never know) it would likely be in the assumptions on mortality/asymptomatic infection/asymptomatic transmission. And I don't think anyone claims Ferguson's or LSHTM's assumptions were wrong based on what was known at the time.
There seems a lot of urge to *blame* at the moment. I don't find this attractive.
--AS
Also, in general if you do come up with something that is valuable, that is when you start to employ people who are software engineers, to well, software engineer.
For the next phase we do need Test and Trace and Isolate.
Why not subcontract the whole shebang to South Korea. I don't think those in Seoul are bothered by where I am, nor want to use me in their big data.
Indeed with a Samsung phone they probably know already...
Pulling a stunt of sticking all your massive project in one mega file will get you a very bad mark even at first year undergrad computer science classes. Before you even get to no documentation.
This is stuff is literally programming 101.
It's why I'm not concerned about that nor the single 15,000 line file a single 15,000 line class would be an issue, a correctly written program in a single file isn't perfect but it's just a bad decision rather than an issue.
We use supermarkets as a top up, maybe once a month, for tinned goods etc. Our food gets delivered by Abel and Cole. Fresh and organic meat, fish, veg, fruit, cheese. Yeh we pay a bit more, but they've never missed a beat during this crisis.
My parents, who have been supermarket stalwarts my whole life, are astounded with the difference switching to their local farm shop which now delivers.
Good food small businesses, farm shops, delis, butchers, bakers etc are going to have a boon after all this has shaken out.
You can have an amazing model but if your assumptions are flawed it will be just a case of garbage in, garbage out.
I worry when people defer too much to a model, any model, without thinking about the assumptions that underpin it.
Of course you would hope software was better written, but it's the model that matters most, not the implementation, even if it it quite poor.
Moreover, for those who do still have to go out to work, getting a large fraction of the traffic off the roads and the bulk of the commuters off the trains will make life so much less stressful, and save millions of hours (and an appreciable number of lives) every year from all those traffic jams, road accidents, broken down trains and "train cancelled due to no driver available" episodes that will no longer happen.
I frankly don't see who loses from all of this except the public transport providers. Especially in the current climate, very few voters will be troubled if they all end up being quietly renationalised and run under something like the TfL model. It's a no brainer.
The errors are not likely to be in the maths though, but rather the assumptions. What is the R number unrestrained? What percentage of people will comply with lockdown? What is the prevalence of the disease in the community? There are too many things that we simply do not know yet.
538 did a very good comic strip on this:
https://fivethirtyeight.com/features/a-comic-strip-tour-of-the-wild-world-of-pandemic-modeling/amp/?__twitter_impression=true
Maybe they’ve all been running fifteen year old code.
Edit: when I said style shouldn't be critisied, I mean in a way that seeks to suggest that the model isn't accurate just because it doesn't use parellelisation or whatever.
I'll get my coat....
Not going to get into the real weeds with the technicalities of machine learning and mathematical modelling, other than to say it is a very different world from 2007.
It isn't just refactoring your code, it is taking advantage of a modern tools and I would bloody hope that in his field algorithms have also evolved.
BBC News - Mass culling for foot-and-mouth 'may be unnecessary'
http://www.bbc.co.uk/news/science-environment-13299666
Q31 in the link
https://www.monmouth.edu/polling-institute/reports/monmouthpoll_us_050620/
If different hardware is producing different results, yeah, you are in a world of pain.
To the earlier point about garbage in garbage out GIGO. Absolutely right and the assumptions and parameters are critical, but you can also get code that itself creates garbage as well as algorithmic approaches that do the same. Models are an exquisite balance of data, assumptions, algorithms, implementations, interpretations and human judgement. A person decides all this. Someone decides the framework, someone picks the data variables, granulairity and derivations. Someone decides what to do with outliers, with nulls, with NaN. Someone creates the decision space, someone picks the language. This is all still a deeply human process.
And yes, I'm sure that the doorstep deliveries of fresh food are lovely, but you would quickly find that these firms were far more hopeless than the supermarkets if everyone rushed to use them. Which, of course, they will not because the vast numbers of people living off the basic state pension, universal credit, zero hours contracts and minimum wage crap jobs cannot aspire to such luxuries.
Niche retailers have their place but they'll remain niche. People who have to work don't have the time to spend wandering around sundry specialist retailers in the manner of a pre-modern housewife doing her daily shopping trip to town. People who aren't in work mostly can't afford to shop in these outlets. And most consumers appreciate the convenience and relatively low prices that a supermarket can offer on a wide variety of goods, all brought together in one place.
This has been about the 8th epidemic so far this century.
Of course, if they were making similar modelling assumptions then the fact the two largely agreed doesn't tell you all that much about the correctness of their conclusions, though it does reduce the probability this is all just down to a bug in one man's code. I've seen plenty of shoddy analytics code in corporate environments for what it's worth but it's often ugly, undocumented, inefficient and despite all that, pretty much works, as accurately as it needed to. Would be rather more reassuring if lockdown was based on transparent, documented, well-written code of course. I think the LSHTM group have been good at sticking their code online throughout the pandemic response (here's their repository) though to far less fanfare than Imperial managed. Which reminds me of something I posted last night...
Being Private Eye everything has to be taken with a certain modicum of salt but even so, their foot and mouth special report is an interesting read. Very telling how the Imperial modelling team at the time (presumably largely or entirely via Roy Anderson, so this isn't necessarily a direct comment on Ferguson himself who was the junior partner) carefully cultivated the media not just to improve the research group's image/standing but to try to "bounce" policy decisions when rivals were suggesting the evidence favoured alternative approaches.
(The eagle-eyed may note Christl Donnelly being quoted a lot in pieces about her colleague Ferguson at present! And recall the Oxford epidemiologist Sunetra Gupta for her criticism of the Imperial team's modelling assumptions earlier in the epidemic, a bit - or rather, a heck of a lot - of history there...)
Doesn't detract from the fact Anderson / Ferguson are high-profile figures who have made very important contributions to the field. But ability to play the media is a handy skill to have if you're involved in policy-determining research.
https://www.theguardian.com/world/2020/may/06/uk-scientists-being-drawn-into-very-unpleasant-political-situation
Every £1 of income (e.g. from one partner working) reduces that by 63p.
I still run old code when it implements something I need, even if it could be refined with modern techniques. It's quicker to do that than refactor and rebugfix. If it needs to run faster, throw money at AWS. Indeed, when it's battle-tested code, I prefer not to fiddle with it. My experience is that each grant is different, and tends to need new code doing something novel rather than a refinement of old code. And unless it's funded, there is neither time nor money to refactor (I rarely touch code myself any more, leaving that for postdocs and students!).
If the government asked my advice (on something within my expertise), I'd most likely run old code. Maybe even 13 year-old code. I myself would be uneasy if my results were then the only thing leading to policy, but if my peers were coming to similar conclusions then I'd be pretty confident that we collectively got the right answers from the inputs we had.
--AS
But instead of just restoring season tickets (traditionally bought by employees comnmuting in), I'm thinking of offering a voucher scheme, whereby employers could buy books of vouchers or their electronic equivalent to give to staff who came in two or three times a week. So businesses that can't yet afford to have their staff all turn up every day could ve encouraged to have half for two days, the other half for another two, or whatever worked best for them
Does anyone have any experience of that approach?
Sometimes I call in.
I would recommend enabling that feature.
--AS
The peak was a close run thing, but the limit is going to be skilled personnel, and medicines. These are very sick patients with multiple organ failures. Dialysis stuff is short too.
These are international issues of supply, not particularly of our own governments making.
When in fact, we know from early January work was being done and those early models used early Chinese data, and it was when more up to date stuff in particular the Italian data, then the 8x capacity horror showed up.
The start of his comments.
https://twitter.com/ID_AA_Carmack/status/1254872368763277313
Note this one.
https://twitter.com/ID_AA_Carmack/status/1254878975459045381