all 9 comments

[–]neolib 2 insightful - 2 fun2 insightful - 1 fun3 insightful - 2 fun -  (0 children)

This is from May 2023, there was a huge thread on HN at the time: https://news.ycombinator.com/item?id=35813322 (2455 points, 1333 comments)

[–]xoenix[S] 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (0 children)

Found via this thread on Reddit.

[–]Canbot 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (5 children)

And yet they still created the most restricted, ideologically corrupt AI of anyone. I guess their plan is to either suppress true open source AI's through regulation and slander, or kick the can down the road and try to compete with those offerings when they emerge by derestricting theirs when that time comes. I think the former will fail because AI is easy, if costly and time consuming, to create and they have yet to create a one world government capable of total control over everyone. And the latter will fail because they have already become a laughing stock and destroyed their own reputation, which will continue to get worse over time. People will not just forget.

[–]xoenix[S] 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (4 children)

The threat for Google is that they may lose their search and advertising monopoly if AI eats search. They can't survive being downgraded to a company that supplies public schools with AI products, bidding against Microsoft, Apple (who may be a dark horse in this race) or whoever.

It doesn't seem likely that they can stop open source though. Heck, even when it comes to the safety layer, the OS ecosystem will probably outpace Google's DEI nightmare team, with much more flexibility.

[–]Canbot 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

outpace Google's DEI nightmare team

I actually think that AI will greatly mitigate the harm of DEI in the short term. That is, at least for the high wealth corporations like google who will be the first to develop AI productivity tools. That is because GPT's do not require intelligence to build, just gobs of data and processing. The AI will be the intelligence that the DEI workforce lacks.

Long term though higher IQ people will become better at utilizing AI and the paradigm will shift. But that won't happen until useful AI becomes available to the public.

[–]Canbot 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (2 children)

You are 100% right that AI will gobble up search revenue. Especially from known propagandists like google. Frankly I am impatiently waiting for the open source AI web crawlers which should be better than search engines right out the gate, but I am not intelligent enough myself to build that. I can just imagine having an app on my phone that uses some trustworthy version of lama where I can prompt it to go do research on whatever topic and it will slowly but surely scower the internet in all languages, through tor nodes, evading ISP blocks, intelligently filtering out bullshit and clickbait, and systematically verify data according to my instructions and then compile it into an a research paper. Even if the process takes weeks to complete it would be life changing.

[–]xoenix[S] 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (1 child)

Let's just appreciate the irony of Microsoft's Github hosting the development of a potential ChatGPT killer.

[–]Canbot 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

Naturally microsoft would want any AI development done on their platform, under their close supervision. But I suspect they would try to secretly sabotage those efforts. So there is a good chance that the GPT killer will actually come from some college club or hacker collective, with a lot of their work done using github, but the secret sauce being developed on their private systems.

Before there is an open source, or 3rd party gpt killer there will be other GPT's that overtake chat gpt, like Apple's thing, and on a commercial level google is also a competitor. So between the tech giants, nvidia, grok, lamma, stable diffusion, claude, mixtral, Mosaic, etc. there is going to be a very dynamic GPT landscape.

[–]iamonlyoneman 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (0 children)

Wait, people are PAYING for a machine to commit mental masturbation for them?