sam0x17

6 ms·

1.
by sam0x17 1mo ago
> GPT 5.6 Sol is the best "vision" model OpenAI ever released I mean I should hope so, as it is also the latest one
2.
by sam0x17 2mo ago
I didn't say fix, I said context. As in "you are using an established thing the wrong way" which is most questions people ask. Edited
3.
by sam0x17 2mo ago
> I can talk to Claude myself OK but can you talk to my specific claude code conversation that has 6 months of context on the mechanism you are struggling with?
4.
by sam0x17 2mo ago
> The only thing a ban would do is protect the American market from further downward price pressure on inference, protecting VC investors in the short term It won't even do this. Streisand effect will probably draw even more attenti
5.
by sam0x17 2mo ago
Mark Twain
6.
by sam0x17 4mo ago
I wonder what kind of result you'd get with floats
7.
by sam0x17 4mo ago
What I like is it's still hard, but you can do things like "prove this using game theory" or "find the optimal value for this and a proof" without having to know game theory or deep math really
8.
by sam0x17 5mo ago
There used to be a lot of research into using deep NNs to train decision trees, which are themselves much less of a black box and can actually be reasoned about. I wonder where that all went?
9.
by sam0x17 5mo ago
Sounds like a vibe-coded feature if I ever heard of one
10.
by sam0x17 5mo ago
It's really gone to shit in the last 2 years
11.
by sam0x17 5mo ago
Very ironic for the billionaire to be openly replacing himself with AI, I suppose he believes his job is easy enough that an LLM can do it, so we definitely don't need him
12.
by sam0x17 5mo ago
If you have no vulnerabilities in your code you have nothing to worry about
13.
by sam0x17 5mo ago
What I mean by "nothing special" is 1 year from now you will say it is an extremely limited model compared to whatever is out then
14.
by sam0x17 5mo ago
It's effectively 2026's version of "Doctors hate this one weird trick!"
15.
by sam0x17 5mo ago
It's all just really genius marketing. In 6 months Mythos will be nothing special, but right now everyone is being manipulated into fearing its release, as a marketing ploy. This is the same reason AI founders perennially worry in publ
16.
by sam0x17 6mo ago
Yeah perfect example, the main thing I _would_ use multiple agents on is optimizing/benchmarking code, but for that you specifically can't use worktree, you need one agent per machine or they'll taint each other's benchm
17.
by sam0x17 6mo ago
The throwing dildos at WNBA games thing https://www.usatoday.com/story/sports/wnba/2025/08/07/wnba-s...
18.
by sam0x17 6mo ago
It's inherently free data. You have to be able to trade against it, and to trade against it you need to see it _before_ you give them money, so they can basically never charge for it. When they sell it to people, it's just lazines
19.
by sam0x17 6mo ago
> It's not some kind of magical money multiplication machine Right, it's not even consistent, the manipulation can go either way, at random, but what is sure is the higher the liquidity, the higher the motivation becomes to tam
20.
by sam0x17 6mo ago
Especially when the "prediction" part isn't even part of the product, it is an accidental byproduct. No one pays for access to the predictions. And they aren't predictions, they are more like "outcome-shaping"
21.
by sam0x17 6mo ago
If anything, the stricter the compiler the better for vibe coding the language
22.
by sam0x17 6mo ago
Polymarket being an incompetent cabal that doesn't deserve to have a license is so passe, so ingrained in the standard recurring narrative that comes up whenever prediction markets are mentioned, and so universally agreed upon at this
23.
by sam0x17 6mo ago
> because it was the only way I could eke out something remotely relevant to the article from your original comment. Remotely relevant? It's not like I'm talking about cats, I'm talking about prediction markets, which is t
24.
by sam0x17 6mo ago
It's also fundamentally something you can't really do properly in crypto. You need a central authority to decide if "thing has happened". There's no way to properly incentivize accuracy in cases where the majority o
25.
by sam0x17 6mo ago
I mean the real purpose is to make money on intelligence-based bets. And what's great is after the fact when everything decrypts you would still get to see the exact timeline of sentiment as it changes, completely undisturbed by the ma
26.
by sam0x17 6mo ago
You're still making a prediction, it's just much less likely to affect the outcome I personally think that entire social value rationale is actually what makes prediction markets dangerous / socially bad Seeing the total liqu
27.
by sam0x17 6mo ago
The problem in the article is sort of boring, this is nothing new, centralized authorities who run prediction markets have been doing a terrible job at wording the predictions and then people get pissed when those poorly written words are p
28.
by sam0x17 6mo ago
You didn't even read my message. At no point do I say the oracle target should be private, that would be ridiculous. Public oracle, private markets on both sides of the issue, until the issue is considered closed, at which point we get
29.
by sam0x17 6mo ago
> I'm not parent commenter, but any example of an event that actually matters? Elections, wars, the economy, healthcare systems, laws, court cases? I think the point is more that if it's already happening at this level of adopt
30.
by sam0x17 6mo ago
many many, like that stuff with the "will someone throw something at X basketball game" and then someone kept doing it, etc

More ›