whatever1

26 ms·

1.
by whatever1 4d ago
It is very, very hard to enforce behavior to an optimization system just with rewards / penalties and no explicit constraints. Which is why in manufacturing we use MPC, not RL (or use them within a system that can outright reject their
2.
by whatever1 6d ago
At some point we will lose track of all the ai discoveries that are worth remembering. Academia with the publication system had a way of retrieving old discoveries and build upon them. If my LLM session found something groundbreaking in bet
3.
by whatever1 6d ago
Better at writing one piece of code, maybe. Better at writing code within a huge system, definitely not. Maybe in the future, but as of Astra, Fable 5.1, the answer is still no.
4.
by whatever1 8d ago
Taxis also save lives. The lives of those who drive drunk.
5.
by whatever1 8d ago
I wish next year for videos too. We really need it for journalism. Otherwise we will not know what is real and what fabricated with ai.
6.
by whatever1 10d ago
I think compression algorithms are ripe for significant improvement with LLMs. It’s an ideal candidate problem you can have in a closed loop evaluation, and you can just let agent try things.
7.
by whatever1 10d ago
I love the fact that 2 days after the ceo surfaced basic usability issues there was still no tangible action plan. It’s insane how much runway a monopoly can buy you.
8.
by whatever1 11d ago
This message would be so badass if you were Ternus himself!
9.
by whatever1 11d ago
Why Apple does not at least provide the specs to the community? So much wasted effort when they can just share a pdf.
10.
by whatever1 12d ago
Yeah in general it’s good advice to not existentially challenge the Germans.
11.
by whatever1 12d ago
Maybe it will be a good incentive to rebuild our global infrastructure without having to go through a world war ?
12.
by whatever1 15d ago
Then the best people quit on a random Tuesday and we have all hands on why managers cannot retain top talent. Nobody believes a single promise of a company anymore. That promo promise after a hard assignment? Yeah fuck that.
13.
by whatever1 19d ago
EU will be the happy island of human bureaucracy standing against the AI Armageddon.
14.
by whatever1 21d ago
`Chinese foreign policy is more dovish than public opinion, especially regarding Taiwan.` Dovish as in it will be annexed in the next decade and not next week?
15.
by whatever1 21d ago
China’s surveillance state seems very successful at managing population emotions.
16.
by whatever1 22d ago
We don’t need my hypotheticals. We saw how well the American bootcamps filled the market with completely trash talent when the market demanded a big number of SDEs. So no, money cannot generate infinite specialized talent overnight as you n
17.
by whatever1 22d ago
Money does not fix the supply problem the parent comment pointed out (at least for fields that need multi year training). Let’s say that tomorrow we need 1M nuclear physicists. How can money help? You still need to wait for the talent pipel
18.
by whatever1 24d ago
It’s a bunch of LLM developed scripts for Linux. Free is too expensive.
19.
by whatever1 26d ago
Typically screen time is very scattered through the day. You cannot read a book in 30” increments.
20.
by whatever1 27d ago
Also if any of the local governors / entrepreneurs cross a line they just disappear to focus on reading. Some of them might reappear after a decade having thoroughly understood the textbooks. For example Jack Ma reappeared after his su
21.
by whatever1 28d ago
Twitter is dead. After the layoffs it required multiple bailouts and pivots to its business. Not even sure what X is about today. Layoffs pushed over the cliff the company. Both due to engineering (content is now trash), and product (lost m
22.
by whatever1 28d ago
Deal. See you here 2031. If the site is still up.
23.
by whatever1 28d ago
Wait two years until we have sufficient churn of senior talent in the teams. Then all of the services will have outages daily. Only the seniors who know their systems are keeping the lights on today by keeping bs commits out. Once they burn
24.
by whatever1 28d ago
Oh you have not played backgammon. You cannot imagine the sheer amount of man-hours we have invested as a species to moving checkers on a piece of wood.
25.
by whatever1 28d ago
Let’s take this argument to the extreme. Why reading mein kampf is healthier than looking at random funny videos? Both are addictive. The one is more dangerous than the other, demonstrably. Don’t get me started on religious books.
26.
by whatever1 28d ago
I get the TikTok addiction, but before that it was fb, before that was the tv, before that the newspaper/magazines, before that the books, before that table games etc I don’t know it seems that regardless of the technology the brain ne
27.
by whatever1 28d ago
I am done with time, extra seconds, missing days, time zones, daylight savings, relativity. Distance from the big bang from now on.
28.
by whatever1 29d ago
The top performers I have worked with will challenge dumb processes, because they can zoom out. Sometimes these dumb processes are there for a reason and you just have to follow them, no questions asked. Example: military. They literally ge
29.
by whatever1 29d ago
This reflects humans. If you need reliability for a clearly defined set of problems you don’t hire a superstar. They will keep poking at the problem, drive it to directions you did not intend to and ultimately they will be worse at the task
30.
by whatever1 1mo ago
You will not believe the amounts of ram I will buy once the whole thing crashes. Like my raspberry pi clone will be running on a terabyte of ram.

More ›