Hacker News new | past | comments | ask | show | jobs | submit
close04 | on: Root shell on Samsung TV
The coincidence is that they took the correct path out of “every possible one” exactly after one of their users took it.

Monkeys on typewriters excuse only works for huge companies who can afford to buy opinions.

FinnKuhn | parent | next
> The coincidence is that they took the correct path out of “every possible one” exactly after one of their users took it.

From what I read online that researcher also used a lot of AI to find this approach, so it might just be the one that is the easiest to find for AI agents and the timing would then only be caused by them working on it around the same time.

HighGoldstein | parent | next
They also took the correct path for hundreds of other problems, where they didn't have a user that already did.
kingstnap | parent
> The coincidence is that they took the correct path out of “every possible one"

My point was explicitly that this didn't happen. I said that they took every possible path and published the one that worked. It's exactly throw shit at the wall and see what sticks.

> exactly after one of their users took it.

All parties agree that what happened is that someone had told someone working at OpenAI that Anthropic employees are working on Navier stokes. This is where the timing comes from.

---

What you are implying is this: OpenAI didn't do the throw shit at the wall approach as outlined in their release blog they actually secretely monitor everyones chat logs and found the idea in Tristan's Chat history and stole and published it.

Or some variant of this. You are only implying what you think is the truth, please state it outright. Don't beat around the bush.

close04 | root | parent
I’ll start with this because it sets the tone for how uninformed your opinion is, thinking you're calling out some conspiracy theory:

> they actually secretely monitor everyones chat logs

Are you for real? They openly monitor everyone’s chat logs. If you had bothered to ask ChatGPT it would literally tell you that they do it. That’s how they know (and yet still fail to report, so the mechanism isn't there for reporting to authorities) that someone is planning a school shooting [0]. They monitor anything and anyone of interest.

Now if “having agents” was enough then we’d just ask for the cure for cancer or proof for every theorem and wait for a few days. Obviously they need guidance to not get stuck in a dead end or have to brute force everything. OAI’s mathematicians weren’t able to provide a good direction so the agents stalled until someone else found the key and only then OAI had their breakthrough.

[0] https://airisk.mit.edu/ai-incidents/1375-openai-allegedly-di...

kingstnap | root | parent
I know they monitor peoples chat logs. I'm on hacknews as much as everyone else and have read that story as well as the ones where they did report people to police based on chat logs. Thats not what I mean by conspiracy because its not conspiritotial. It's literally stated on the help page [0].

I'm not uninformed on this. I actually wasted a lot of time on it since my interests intersect at this. I've watched the numberphile interviews, read the drama, various takes etc.

What I'm saying is that its highly unlikely that in this case that Sébastien Bubeck from OpenAi decided to steal some critical idea from Buckmasters chat and feed it "Bell" the unreleased model that produced the solution.

> Obviously they need guidance to not get stuck in a dead end or have to brute force everything

Dude they did brute force everything. This is what I keep trying to say. They said they spent 130 billion output tokens.

This is quite literally like 500,000 books worth of text.

[0] https://help.openai.com/en/articles/6783457-what-is-chatgpt

loading story #50235281