highfrequency

4 ms·

1.
by highfrequency 10d ago
Preposterous, perhaps - but if the role-play is convincing enough for large groups of people, it could start to have impact on human decision-making. The crowds have been swayed by much more preposterous narratives. I believe Suleyman is ar
2.
by highfrequency 10d ago
From the actual essay ( https://mustafa-suleyman.ai/a-warning-about-model-welfare ): > They go on to write – speaking directly to Claude – that “questions about Claude’s moral status, welfare, and consciousness remain deep
3.
by highfrequency 10d ago
> many tasks are covered in the training data; but even small perturbations within a covered class of task result in outright failure or reward hacking. Is this true? Could you cite an example of a simple prompt that GPT 6 / Fable 5
4.
Anthropic team lead: AI >10% to kill all humans in next decade(twitter.com)
3 points by highfrequency 12d ago|0 comments
5.
by highfrequency 14d ago
Let's be real - if OpenAI and Anthropic did not accelerate the scaling and monetization of LLMs as fast as they possibly could from 2019 to 2024, it would be many, many years until China/Russia/North Korea spontaneously launc
6.
by highfrequency 15d ago
What providers do you find most reliable? And is that general across models?
7.
Defeat in Detail(en.wikipedia.org)
2 points by highfrequency 16d ago|0 comments
8.
by highfrequency 18d ago
This would be a fairly insane breach of trust and common sense if true; the chain-of-thought / reasoning trace is, from an information perspective, close to a superset of the prompt and model response.
9.
by highfrequency 18d ago
> If they could declare with certainty that Buckminster's and Alpoge's usage data had been totally excluded from training, would that set a worse precedent and reflect poorly on their de-identification process (and data access
10.
by highfrequency 18d ago
From OpenAI: > While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models This is the crux of it. If Tristan's work and insights were not used to train OpenAI models
11.
by highfrequency 18d ago
> While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models This is the crux of it. If Tristan's work and insights were not used to train OpenAI models, then this j
12.
Bill Gates tries to install MovieMaker (2003)(techemails.com)
436 points by highfrequency 19d ago|295 comments
13.
by highfrequency 22d ago
Sure, politics and economics are involved - but why can’t technology also play an important role? None of this AI political economy stuff was relevant in 2015 because necessary technological breakthroughs had not yet been made.
14.
How the iPhone blew up the Wireless Industry(wired.com)
3 points by highfrequency 22d ago|0 comments
15.
by highfrequency 27d ago
To clarify, is the TLDR that state of the art models were prompted to cheat / exploit their environment and they did so successfully? Or did OpenAI prompt the models to not cheat and they did anyway? Surprisingly hard to get a clear
16.
by highfrequency 29d ago
> Go on breaking the law until you get a favorable judge/Congress. Nobody is going to jail (we don't really do that anymore), you might pay a fine. Binance, FTX, BitMex, Silk Road founders all went to jail - it’s a huge coin fl
17.
by highfrequency 1mo ago
> the "token spewer" work. being ultra responsive, pushing the ball forward across dozens of different fronts... ~95% of the work he does falls into bucket 2. It's hopping on calls. Nudging people. Blocking and tackling. T
18.
by highfrequency 1mo ago
It is sometimes the opposite - a large number of things makes the system easier to predict and reason about (statistics, behavior of gases etc).
19.
The Blow Up Artist (Victor Niederhoffer)(newyorker.com)
1 points by highfrequency 1mo ago|0 comments
20.
Agency Costs of Free Cash Flow, Corporate Finance and Takeovers [pdf](liuyanecon.com)
1 points by highfrequency 1mo ago|0 comments
21.
Crash of 1987: Leo Melamed and the delayed wire(leomelamed.com)
2 points by highfrequency 1mo ago|0 comments
22.
by highfrequency 1mo ago
> But it has one property much of my previous work lacked: reality gets a vote. Code fails. The parsers disagree. And the data can refuse the story I want to tell. In this project and elsewhere, I am now trying to judge work by what surv
23.
by highfrequency 1mo ago
Terence Tao's quote about AI's math proofs is relatable outside of pure math: "the writing very often dwells at length on trivialities while passing briefly through — or even actively obscuring — the most interesting and nove
24.
The SEC is no longer independent(bloomberg.com)
10 points by highfrequency 1mo ago|2 comments
25.
by highfrequency 1mo ago
Very cool - to be clear, are both your diet composition and calorie intake similar before vs. after?
26.
by highfrequency 2mo ago
Perhaps customers choosing your product for irrational philosophical reasons is exactly how you'd want to position your product if you're a business. When it comes to margins, the only thing better than a price-insensitive custo
27.
Guy Spier's 2025 Letter to Partners (after diagnosis)(aqfd.docsend.com)
3 points by highfrequency 2mo ago|1 comments
28.
Atypical Combinations and Scientific Impact [pdf](kellogg.northwestern.edu)
1 points by highfrequency 2mo ago|0 comments
29.
by highfrequency 2mo ago
Are you aware of her APOE3/APOE4 status?
30.
Elon Musk and the Infinity Rebuy(davekarpf.substack.com)
9 points by highfrequency 2mo ago|0 comments

More ›