Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Andrej Karpathy Departs OpenAI (theinformation.com)
182 points by sstojanov on Feb 14, 2024 | hide | past | favorite | 37 comments



It would be helpful if the title of that discussion indicated that "I" is "Andrej Karpathy".

This discussion doesn't suffer from that defect.


The name is in the domain to the right of the title.


This has got to be the most user hostile website I have ever seen, by quite a margin. Wow. It’s only missing the spinning roulette when you go to exit the site.


Karpathy just tweeted about it [0]:

> Hi everyone yes, I left OpenAI yesterday. First of all nothing "happened" and it’s not a result of any particular event, issue or drama (but please keep the conspiracy theories coming as they are highly entertaining :)). Actually, being at OpenAI over the last ~year has been really great - the team is really strong, the people are wonderful, and the roadmap is very exciting, and I think we all have a lot to look forward to. My immediate plan is to work on my personal projects and see what happens. Those of you who’ve followed me for a while may have a sense for what that might look like ;) Cheers

[0]: https://twitter.com/karpathy/status/1757600075281547344



The article is behind paywall. Anyone care to give link without paywall?


Andrej Karpathy, one of the founding members of OpenAI, has left the company, a spokesperson confirmed.

Karpathy, a prominent artificial intelligence researcher, was developing a product he has described as an AI assistant and worked closely with the company’s research chief, Bob McGrew. While ChatGPT has been a hit with consumers, OpenAI wants to launch software that can automate complex computer-based tasks, like filling out expense reports and entering them in accounting software, The Information reported last week.

Karpathy rejoined OpenAI last year after spending five years at Tesla, where he oversaw development of its Autopilot semi-automated driving software. His departure from OpenAI comes almost exactly one year since he said on X that he was returning to the company.

Karpathy couldn’t immediately be reached for comment. “Andrej is departing to pursue personal projects. We are deeply grateful for his contributions and wish him the best,” OpenAI spokesperson Kayla Wood said in a statement. “His responsibilities have transitioned to a senior researcher who worked closely alongside Andrej.”

Karpathy is the first high-profile departure at OpenAI since several senior staff resigned in the wake of CEO Sam Altman’s ouster by the board of OpenAI’s nonprofit parent. The staffers returned to the company alongside Altman after the board reversed its decision. Despite the drama, OpenAI’s business growth continued.

Karpathy has been a public face of OpenAI through podcasts, and he posts frequently on X. He has described large language models, the conversational AI that powers ChatGPT, as a kind of operating system because of its ability to retrieve files, write code and run programs, and understand audio, images and human commands.


Bob McGrew from Palantir is OpenAI's research chief? That's one hell of a move up for him!


Not sure how I feel about this. I think just telling the gist of the article is fine, it’s like me talking to a friend, but sharing the article word to word seems like stealing.


It's kind of a protest against the unethical practice that sites like this often do where they let search engines see their entire article for free just to get indexed and ranked highly, only to lock out visitors who click on those links. It's deceptive and manipulative just to get clicks.

gary_0's reply below is a good way to frame the deception: show the search engine one thing, show the user something else.


Before they switched to pro-evil, Google used to ban you from their index if you did that (show them one thing, and human visitors something else).


Which itself had funny consequences. Expert Sexchange used to show you a page with a payment dialog up top but you could just scroll down to the content.


I want to believe this is an intentional misspelling.



I agree that we should criticise two-faced paywalls. We can do this by not linking to paywalled articles in the first place. After all, this is just reporting of a newsworthy event, other publishers exist — or will exist within minutes. (And if there aren't at least two good sources of reporting, perhaps we shouldn't assume it's true anyway?)

It's important to remember that respect for the fundamentals of copyright is a prerequisite for respecting the GNU GPL and other such open source licenses.


Should he have rephrased the info instead?


Yeah, run it through ChatGPT to summarize it and get some truly original content


I don't see how this is any different than posting an archive link.

In this case, theinformation.com articles are still paywalled in archive links. So I just pasted the contents.


HN people always post archive link to pass paywall. It's explicit piracy. But since copyright is evil (when it's not mine) and content creators aren't worth their salt (if they're not me) so it's okay to do that.


I always thought some of the people posting the archive links were shills for the people running that site (it’s monetized with ads) — brilliant strategy.


I'd be disappointed to learn that anyone who's been longer on HN than about five minutes wasn't using an adblocker.

That said, this is a valuable service. If I were ever willing to consider turning off an adblocker anywhere, this would be where I'd start.


Ah, but that’s not the only way —- backlinks, people copying the link and sharing it on social media, and mentioning the site all grow traffic immensely, and some of that traffic won’t have ad blockers :)

IIRC one of the sites still has a “buy me a coffee” link, as if the owner needs it.


Well there is more content than eyeballs at this point, an exponentially rising gap between the two you could say. So according to economic theory when production outpaces consumption what happens to the value and price of whats produced?


>But since copyright is evil (when it's not mine) and content creators aren't worth their salt (if they're not me) so it's okay to do that.

This is a disingenuous generalization.


We can't have a discussion on a link behind a paywall. That's what the comments section inside your paywall is for.

If a link is posted here and requires a subscription, how does that further the conversation you are intending by posting the link?

I want real humans paid for their efforts! If that is the primary concern then we shouldn't be posting non-public links into a public forum for discussion?


As a journalist, I defend the habit of sharing archive links. Paywalls are a lazy and detrimental solution to a larger financial problem for media organizations. They limit the access to information to a specific subset of people, and let misinformation run wild and free.

So yes, sharing the article is important in this case. If there was, say, a 300$ per year global subscription to a 100 of publications, I'm also sure 90% of hacker news would buy it.


> If there was, say, a 300$ per year global subscription to a 100 of publications, I'm also sure 90% of hacker news would buy it.

No one is going to pay $300/yr to read news articles online.

But even people do that, by your logic it's just limiting the access to information to a specific group of those who are able and willing to pay $300/yr to read news.


I assure you that many people spend $300 annually on news and journalism.


This kind of already exists with Apple News+. Not sure how well that is going but if anyone has the reach to make this work it would be Apple.

Personally I am for anything that can decouple advertising from reporting. There are just too many perverse incentives that it creates.


That site is straight out a hostile asshole design.


Ok, but please don't post like this to Hacker News. We're trying for something else here.

https://news.ycombinator.com/newsguidelines.html


I will keep that in mind, sorry.


It's quite funny looking with JavaScript disabled.


Paywall and archive doesn't work with this website. Can anyone with access share more details? Is this baseless speculation or has it been confirmed? Do we know what he plans to do going forward?


I have the paywall as well, but in the opening sentence they mention they confirmed it with a spokesperson at the company. Sounds pretty legit.


I'm not saying the article is made up, but if it was, the opening sentence could still say they confirmed it with a spokesperson at the company.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: