<< Higher ups may want it done, but probably only if it can be done immediately and doesn't cost a lot of money.
Since we are talking evil, I suppose money is a good place to start ( being root of it and all that ). So motivation would likely be there, but I am willing to accept your qualification of 'yes, but don't spend a lot'.
Lets look at some of that potential cost structure. The analytics part these days is not exactly expensive. Hell, HN just yesterday had a story about $4.80 GPU time being used to do some reinforcement learning to find 'best HN post' on 9gb of HN data[1]. That used to be a little require more time and be more labor intensive. Edit: Yes, what we are discussing would naturally go beyond $4.80, but in terms of bang for buck it is hard to find a better time than now.
One could reasonably argue getting the right people ( right experience and knowledge ) could be prohibitive in terms of cost, but.. if you are already an insurance company, it is not exactly impossible that you already hired people with applicable experience, knowledge and skill. And if you are already paying them, maybe this little offshoot project could be sold to them as a great advancement opportunity. And if they took it too far? Well, no one told them to go overboard. After all, we at <company D> have a strict ethics policy.
If that is the case, two big pieces of the cost structure are either negligible or already part of the annual budget.
>if you are already an insurance company, it is not exactly impossible that you already hired people with applicable experience, knowledge and skill.
my experience with large companies is that there is already allocation of those resources somewhere, with lots of managers and such, I think it would be a major thing to move people around or to hire new people.
Sure it's nice to believe Evil is working agile, but really it just says its adapted some agile methods and its super slow as per the usual.
Since we are talking evil, I suppose money is a good place to start ( being root of it and all that ). So motivation would likely be there, but I am willing to accept your qualification of 'yes, but don't spend a lot'.
Lets look at some of that potential cost structure. The analytics part these days is not exactly expensive. Hell, HN just yesterday had a story about $4.80 GPU time being used to do some reinforcement learning to find 'best HN post' on 9gb of HN data[1]. That used to be a little require more time and be more labor intensive. Edit: Yes, what we are discussing would naturally go beyond $4.80, but in terms of bang for buck it is hard to find a better time than now.
One could reasonably argue getting the right people ( right experience and knowledge ) could be prohibitive in terms of cost, but.. if you are already an insurance company, it is not exactly impossible that you already hired people with applicable experience, knowledge and skill. And if you are already paying them, maybe this little offshoot project could be sold to them as a great advancement opportunity. And if they took it too far? Well, no one told them to go overboard. After all, we at <company D> have a strict ethics policy.
If that is the case, two big pieces of the cost structure are either negligible or already part of the annual budget.
FWIW, I want to think you are right.
[1]https://openpipe.ai/blog/hacker-news-rlhf-part-1