Hacker News Reader

Comment by itchy_spider44 | Hacker News Reader

Ilya is the stereotypical genius mind that is extremely passionate yet disconnected from the real world. He got way too worked up about abstract issues, failed to see the bigger picture and had a meltdown that other board members took seriously because he's a cofounder. He is instrumental in the research but he shouldn't be running the business.

View on HN

> He is instrumental in the research but he shouldn't be running the business.

To push back on this a bit. If two yet unknown people, "an Altman" and "an Ilya", both applied to YC to start a company that builds and sells AI models, guess who would get funded. Not the guy who can't build AI models.

I find it bizarre that the guy who can build is suddenly the villain-nerd who can't be trusted, and the salesman is the hero, in this community.

There has been a huge cultural shift in tech towards devaluing builders and (excessively) emphasising the roles of operations and sales.

In reality you need all of them, and they are all separate talents, but things are clearly unbalanced.

It's how whole industrial sectors become enshittified.

The word enshittified has officially been enshittified.

Enshittiception

Right, I’ve noticed this trend too. It sucks and I find myself frustrated by how easily people go along with smooth talking sales/business types and allow them to take over positions of power. The truth is that builders can exist and build amazing products without the sales guys, but the sales guys would have absolutely nothing without the builders. I’d even go a step further and say that the best products are built by solo or small teams of impassioned builders and as soon as sales/business types get involved things start tending towards enshittification.

IMHE at YC, I'd put my money on whichever talked a more monetizable game.

They'll definitely both get funded.

Is there any way he can even stay if Altman returns? Seems like one of them is out and making their own AI venture soon.

If Steve can make him stay and remain productive, he's a real business genius

He could possibly stay in a pure Chief Scientist role while abdicating his board seat. But if I were a CEO, I'd have a hard time trusting a C-level role to someone whose vision is diametrically opposed to my own.

He won't. In a company, you want all employees to share the same vision. Ylia and Altman clearly don't. There is no point pretending

Make him the same offer he made Brockman - you can stay on as an employee but you're off the board.

That will keep him employed for the three nanoseconds until he gets an offer from a competitor.

He’d probably be better off going to Meta or HuggingFace and working on getting open source as close to OpenAI’s offerings as possible. I expect that real innovation (vs. commercialization) is now fully dead at OpenAI, with them instead focusing on ROÍ for Microsoft.

Ilya is not a champion of open source: "We were wrong. Flat out, we were wrong. If you believe, as we do, that at some point, AI — AGI — is going to be extremely, unbelievably potent, then it just does not make sense to open-source. It is a bad idea... I fully expect that in a few years it’s going to be completely obvious to everyone that open-sourcing AI is just not wise."

https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-lau...

He said in an interview 2 weeks ago that below a certain capability threshold it is beneficial and good to open source models —- but once you cross that threshold it is a bad idea.

The example he gave is a model that could independently do science.

He's opposed to making powerful models open source, unfortunately.

If "bigger picture" means making more money with less safety concerns, then I think he is totally right to oppose that.

He seems to have gone the way of Eliezer Yudkowsky (but with actual technical chops).

For a very long time I thought Eliezer was one of the least likeable people on the internet, which should be a pretty high mountain to climb. But watching him on the Lex Friedman podcast was really interesting, and putting a face to the writing helped quite a bit in humanizing him. He's obviously neuro-divergent, and obviously very flawed in a number of personality traits, but he's quite intelligent and completely obsessed for more than a decade on a specific topic. I wouldn't just outright dismiss that.

Reply on news.ycombinator.com