3 ms·
> Plans should be measured in decades, execution should be measured in weeks. This makes no sense in the domain they are on. Very few. significant things they
by mianos 3y ago
> Plans should be measured in decades, execution should be measured in weeks.
This makes no sense in the domain they are on. Very few. significant things they do takes weeks. Even planning for a learning run takes months.
There are a lot if things that can be split into just a few weeks for some contrite sprint plan. There are others that take longer before realising intermediate results.
- aeternum 3y agoOne of the key things OpenAI does is very small scale model experiments. You can iterate a lot faster if your training run takes on the order of minutes. So vet ideas there, only allow proven ideas to be part of the longer training run. ML model training has become the new "my code's compiling" except now it takes weeks in many cases. The train then chill mindset that some ML engineers have is not they way to make rapid progress.