vessenes

5 ms·

1.
by vessenes 4d ago
Sorry, I don't understand what this means. What does it mean?
2.
by vessenes 4d ago
I think that implies they're seeing unusual new subscription demand, yes? That's how I'd read it, not least because I worked through four resets this week on Astra, which is un unbelievable amount more inference than I'v
3.
by vessenes 4d ago
Prices are moving inference to highly profitable, oAI seems to have solved their training problems, and they're currently competing nicely with Anthropic on the coding side. I'd wait, too -- why fight this stuff out in public when
4.
by vessenes 4d ago
“We” does a lot of work here. You keep using that word. I don’t think that word means what you think it means.
5.
by vessenes 6d ago
Shenanigans is an inaccurate word; it implies underhanded behavior that's hidden / concealed. I think "to act so aggressively" is more balanced. Here's my answer: If you think we're getting to AGI in the next 9
6.
by vessenes 6d ago
Yep, it's possible. But, we have literally no idea how much a set of prompts would impact training as far as general usefulness. I don't think we even know if Tristan's said he allowed training on his prompting or not. This i
7.
by vessenes 7d ago
This is speculation right now. The idea would be that if someone used the product and granted training rights, which is the default for many subscription levels, then some knowledge would have been imparted into the general weights of the n
8.
by vessenes 7d ago
Thanks! It's just infra, so you could do what you wanted. The clients include a safety reminder, and the datastructures include "unsafe" in the name of the inputs passed around, but that's just a little hygiene. There re
9.
by vessenes 8d ago
I built what I think is a pretty good library and set of tools for this earlier this year -- https://github.com/corpollc/qntm (or `uvx qntm --help`); it includes a cli, python and typescript libraries, and works out of
10.
by vessenes 8d ago
Not arguing against open science - it's super valuable. I'm saying that pearl clutching by people reading Tao isn't useful, because it misses some long history which tells us that this kind of science has been seen as fundame
11.
by vessenes 8d ago
That’s not untrue. But it’s also a misstatement of mathematical history. Many leading mathematicians historically have been highly competitive — Gauss comes to mind. Woe betide the lesser intellect that sent Gauss some ideas. The Newton Lei
12.
by vessenes 8d ago
If those researchers did not opt out then training data might go in. I think it’s a courteous acknowledgement; as was reaching out and examining the direction of proofs themselves. At stake here is a particular mathematician dynamic - ego,
13.
by vessenes 10d ago
I can’t read your original comment, but when the chief scientist of a company that just released six month old software that can use Kicad on your computer to design a circuit board and have it created and shipped to you tells you he thinks
14.
by vessenes 10d ago
The situations aren’t equivalent - luckily in my opinion because the stakes with nuclear are much higher. von Neumann constructed a multinational game theory approach appropriate for weapons. AGI is a much harder problem to corral because t
15.
by vessenes 10d ago
I’m like a 2(.5?) there - I don’t think ASI will care about my kids better than I will for some definitions of better, for instance, and I feel very fuzzy and vague about what actual differences in qualia between me and ASI would yield in t
16.
by vessenes 10d ago
No. These parties are composed of people who most definitely think this way, and therefore will have distinct goals and interests when presented with opportunities. That’s reality quite aside from how a game theorist assesses the situation.
17.
by vessenes 10d ago
This is a good essay, and makes me hopeful. I’m on the record saying that it is extremely dangerous to slow down because the race for AGI is a zero-trust game — defections pay - and combined with a compounding returns model on defection,
18.
by vessenes 10d ago
Anthropic’s Mechinterp did some very fine work on this. TLDR - you can; you train a decoder on neuralese to english and then add a loss function for a roundtrip of english -> neuralese -> english (or possibly n -> e -> n? I don’
19.
by vessenes 12d ago
Looked to me like it missed the chain though.
20.
by vessenes 12d ago
Those tend to be under so-called ‘max’ modes, or ‘ultra’ - where you scale up inference time compute and then choose amongst your answers. Astra is new weights.
21.
by vessenes 12d ago
It’s already running in Hawthorne, CA
22.
by vessenes 12d ago
True, .. and. In this case, the original proof is considered rigorously checked, so finding a bug in the kernel would be nice to know about, but in my opinion would not take away from the accomplishment (FLT in lean using agents) nor the ma
23.
by vessenes 13d ago
I'd like to note that we should remember a formalized Lean proof does have value in that it enters the pantheon of true things other Lean proofs can rely on. Agreed that for the humans, descriptions and being able to 'grok' t
24.
by vessenes 13d ago
Pretty efficiently, apparently, since it saturated ARC-AGI-3 in half of the predicted time, and according to the Chollet blog post on the fly created dense DSLs to describe and analyze individual games.
25.
by vessenes 14d ago
Weave robotics (no affiliation) claims they are folding over 1,000 lbs of laundry a week in homes right now: https://www.weaverobotics.com . I have no idea how well these work.
26.
by vessenes 14d ago
Ooh interesting. Got a link? I guess you’d want to find out f that’s a weird motion tic falling out of rl training or the much simpler explanation. I thought figure was publicly clear the robots were operating autonomously.
27.
by vessenes 14d ago
I guess the VLA proposal is that frontier models will keep extending the planning and action horizon and that the action models will keep improving simultaneously leading to a moment when they suddenly get useful. This seems likely to me, b
28.
by vessenes 14d ago
If you tour SpaceX’s hawthorn factory they will tell you: “raw materials come in, rockets go out.” Spacex manufacturing had a mix of robots and humans manufacturing together on the floor the last time I was there, like six years ago. But it
29.
by vessenes 14d ago
This is a recent essay on robotic workers that does not mention Figure’s 24+ hour livestream of a robot doing package management on a live assembly line. The real reason it’s hard to assess where we’re at with industrial robots is that some
30.
by vessenes 14d ago
thank you!

More ›