Mr. Macron
This regime has been meddling and terrorizing the region for 47 years.
If you can’t do anything stay quiet
I forgot, France brought them to Iran
@mattpocockuk@orcdev@steipete Ralphs loops does consume a lot more tokens. I do the test from time to time and my limits are consumed much more quickly. Like much more quickly.
Also having some kind of implementation plan allows to use models that are much much less expensive for the same kind of output
@orcdev@steipete I find this kind of "look at these devs on Mount Olympus" takes really dispiriting. Framing it this way gives you a free hit without needing to address any substance.
you'll often see hot takes from famous AI people that just don't apply to regular devs
one recent hot take from @steipete and @mattpocockuk :
"I don't use plan mode", and a regular dev thinks ZOMG
until you realize those guys have:
- infinite tokens
- 100 agents running in ralph loops
- entire workflows auto iterating
of course they don't need plan mode
meanwhile regular dev needs plan mode, because regular dev is on a $20 subscription, and regular dev cannot burn 5b tokens a day, regular dev says hello and burns 4% of his session
different game, different rules
@yacineMTB@DominionSoc Canada must create a new public holiday, the sex day. You can only take it if you are going to have sex and the purpose is to have a kid.
Some kind of easter day I guess
All of Canada's population growth comes from immigration.
In terms of domestic births, 42% are to foreign born mothers. The large majority are to first, second, third generation immigrants.
Amongst the youngest generation, Canadians are already a minority.
This transformation of our society is not acceptable.
Canada needs Remigration.
The West will be stronger if NATO is dissolved, because only then will Europe take defense seriously. (With apologies to Poland and the Baltics, who don't deserve this.)
Is autoresearch really better than classic hyperparameter tuning?
We did experiments comparing Optuna & autoresearch.
Autoresearch converges faster, is more cost-efficient, and even generalizes better: 🧵(1/6)
L'élite scientifique française qui parle de l'IA.
On en est là. La technologie la plus impactante de l'histoire de l'humanité. Celle qui va potentiellement résoudre la majorité des problèmes qui paralysent l'espèce depuis des millénaires, la santé, l'énergie, l'éducation, la recherche fondamentale. Une technologie qui sera rétrospectivement vue comme plus importante que l'électricité, internet et l'imprimerie combinés.
Quel naufrage.
TLDR, they're lying.
Or they're not doing anything real. Or they're letting their agents commit mistakes.
Or all of the above.
I've built three different features at once in work trees but the quality noticeably goes way down and my real person / human-in-the-loop reviewer regularly has to do more work to fix what the agents screwed up.
The problem is two fold.
First, you can't context switch this much on hard problems and so your personal problem solving drops dramatically.
Second, these models still make massive mistakes and dumb decisions and if you are not watching them closely, more slips through.
And no, an agent overlord/orchestrator does not fix this because the model itself is stupid and part of the problem. It's a bandaid.