fwlr

5 ms·

1.
by fwlr 3d ago
I assume he’s referring to videos like this? https://youtu.be/gEDaCIDvj6I If so… Very poor choice of metaphor. Nothing resembles AI coding quite so much as the mystical master waving his hands in karate-style movements, see
2.
by fwlr 3d ago
It’s not circular because we put a little bit of money in, and a lot of money comes back Ah yes, it’s not circular, because it’s a pyramid.
3.
by fwlr 4d ago
Conversely, I think mathematics should retain this aspect of “number-poetry”. Consider a mathematical pursuit that lacks number-poetry but retains other features such as attribution, calculation, and puzzle-solving: I think of competitions
4.
by fwlr 4d ago
Your simpler model of the mechanism would seem to suggest the very same action that the article’s more complicated model suggests, viz. find a better training method than reinforcement learning.
5.
by fwlr 4d ago
“A race to the bottom, spurred by commercial incentives, can make [AI] risks more acute. … We have sought […] to create a race to the top .” And then he sketches out a plan for slowing the pace of the race, without changing its destination
6.
by fwlr 4d ago
A more accurate title would be “Farmers Aren’t Bought”, as in “bought by John Deere”, like this author is.
7.
by fwlr 5d ago
Man, it would be awesome if I could tell my boss “No, I don’t want to go to prison for 20 years” when they ask me if some magical AI thing can be done.
8.
by fwlr 5d ago
Nick Bostrom: We could imagine, as an extreme case, a technologically highly advanced society, containing many complex structures, some of them far more intricate and intelligent than anything that exists on the planet today – a societ
9.
by fwlr 5d ago
In internet culture there’s this phrase “Hydrogen Bomb vs Coughing Baby”, meant to highlight the absurd power difference between two combatants. In almost any scenario even tangentially involving mathematics, twenty-five Fields medallists u
10.
by fwlr 5d ago
The relentless progress towards saturation of benchmarks is, I suspect, at least partly a similar story. Whatever holdout questions are used to evaluate GPT-x will be in the GPT-x conversation logs, and therefore in the training set for GPT
11.
by fwlr 5d ago
This makes sense. “Don’t try to penny-pinch your employees” is a lesson most managers learn eventually, and I guess agent-orchestrators will have to learn it too.
12.
by fwlr 6d ago
I said “very quickly find enough certainty” to suggest hypothetical situations like “someone searches the conversation logs, confirms for themselves the solution is present, then shares the confidence gained from this knowledge without expl
13.
by fwlr 6d ago
I think you’re overlooking what I’m implying here. It’s not that they knew contamination was possible but they went ahead anyway. To spell it out just a little bit more: learning the answer might be in model X’s training data made them beli
14.
by fwlr 7d ago
It is suspicious that OpenAI decided to generate 300 billion output tokens from a model still in training, right after learning there was a credible chance that a major math proof was in that model’s training data. Obviously there are reaso
15.
by fwlr 7d ago
For my money, the most “AI does my research” mistake I’ve seen in this whole mess is someone saying “the original researchers did not solve Navier-Stokes, while OpenAI did”. It has all the hallmarks of a classic AI blunder - partially eli
16.
by fwlr 7d ago
You might be interested in this paper, rigorously tackling the sorts of questions you raise: https://selfawaresystems.com/wp-content/uploads/2008/01/ai_d...
17.
by fwlr 7d ago
The wet ball is a convenient source of mass and energy to bootstrap the sphere. The fact that extracting those resources changes certain parameters of the wet ball to values that humans no longer find compatible is merely incidental. So it
18.
by fwlr 8d ago
I intimately recognize the familiar structure of “notice something odd when operating software, pull on thread and track down the offending line of code, experience righteous indignation at the bad engineering revealed underneath”. Where it
19.
by fwlr 8d ago
Open math problems, yes; also open source code, art, literature, and everything else as well. AI is a machine for turning commons into tragedies.
20.
by fwlr 8d ago
It’s a pity they had Astra do the writeup. I was curious to see how “GPT7” writes.
21.
by fwlr 8d ago
>The paper is not about implementing a general attack that works on all trusting-trust mitigations. Then why have section 7.3 'Bootstrappable Builds', that briefly surveys a type of trusting-trust mitigation, and concludes that
22.
by fwlr 9d ago
I also think they’ve shortchanged the bootstrap build mitigation. They write: A utility such as strip sustains the attack whenever a tampered instance participates in the rebuild, even after the compiler seed shrinks to a few hundred
23.
by fwlr 9d ago
The attack doesn’t use AI, but an AI could use the attack. (and the researchers definitely used AI to write the paper!)
24.
by fwlr 9d ago
There is so much competition among frontier labs, but it is a competition to see who can burn the biggest pile of cash. Once your competitors flame out, they are gone, and you are free to raise prices. Think WalMart/Uber tactics.
25.
by fwlr 9d ago
Unix pipes are easy, and they are a very good abstraction, and their choice of abstraction boundaries is superlative, but they are not simple. (I have more than once seen a colleague implementing a ring buffer for feeding data into a Unix p
26.
by fwlr 9d ago
A pox upon the editor who gave this piece such a poor title. This is an economic critique of advertising, it is notably dispassionate, and in fact carefully contains precisely zero moralizing whatsoever.
27.
by fwlr 9d ago
The examples in the essay are perhaps not specific enough - they do gesture in the direction of the author’s point, but they also admit other valid interpretations like your own. I think the Rich Hickey talk “Simple Made Easy”, which this p
28.
by fwlr 10d ago
It’s a cynical joke, imagine a person having a personal computer, but in the responsibility-divesting way that corporations have “special purpose vehicle” debt-holding “technically a legally separate entity to us!” shell corporations.
29.
by fwlr 10d ago
The combined intellectual output of the field, across books, lectures, blog posts, verbal discussion in professional and personal capacities (many of them are friends or acquaintances), for about a decade. It’s just not possible to neatly c
30.
by fwlr 10d ago
I have similar feelings, you’re not alone (nor alone in being saddened by it). Things seem to me to be more or less playing out the way I interpreted existential risk researchers saying it would, so there’s a strange relief in knowing the g

More ›