Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The next chapter of the Microsoft–OpenAI partnership

openai.com

Post reply on HN
Print view View original

251–260 of 496 posts

Re: The next chapter of the Microsoft–OpenAI partnership

#251

Post by fidotron »

philipwhiuk wrote: post #79

> Once AGI is declared by OpenAI, that declaration will now be verified by an independent expert panel. > Microsoft’s IP rights for both models and products are extended through 2032 and now includes models post-AGI, with appropriate safety guardrails. Does anyone really think we are close to AGI? I mean honestly?

5-15 years.

The key steps will be going beyond just the neural network and blurring the line between training and inference until it is removed. (Those two ideas are closely related).

Pretending this isn't going to happen is appealing to some metaphysical explanation for the existence of human intelligence.

Re: The next chapter of the Microsoft–OpenAI partnership

#253

Post by wholinator2 »

nl wrote: post #178
no_wizard wrote: post #96

Earlier quoted context omitted.

They’ll devalue the term into something that makes it so. The common conception of it however, no I don’t believe we are anywhere close to it. It’s no different than how they moved the goalpost on the definition of AI at the start of this boom cycle

> they moved the goalpost on the definition of AI at the start of this boom cycle Who is this "they" you speak of? It's true the definition has changed, but not in the direction you seem to think. Before this boom cycle the standard for "AI" was the Turing test. There is no doubt we have comprehensively passed that now.

The turing test point is actually very interesting, because it's testing whether you can tell you're talking to a computer or a person. When Chatgpt3 came out we all declared that test utterly destroyed. But now that we've had time to become accustomed and learn the standard syntax, phraseology, and vocabulary of the gpt's, I've started to be able to detect the AI's again. If humanity becomes completely accustomed to the way AI talks to be able to distinguish it, do we re enter the failed turing test era? Can the turing test only be passed in finite intervals, after which we learn to distinguish it again? I think it can eventually get there, and that the people who can detect the difference becomes a smaller and smaller subset. But who's to say what the zeitgeist on AI will be in a decade

Re: The next chapter of the Microsoft–OpenAI partnership

#254

Post by notatoad »

butler533 wrote: post #247

Why do none of OpenAI announcements have an author attributed to them? Are people that ashamed of working there, they don't even want to attach their name to the work? I guess I would be, too.

because they're corporate PR statements drafted by a team, and corporate press releases don't normally have an author byline

Re: The next chapter of the Microsoft–OpenAI partnership

#255

Post by whynotminot »

alterom wrote: post #249
whynotminot wrote: post #240

Earlier quoted context omitted.

I had a friend who had his Tesla drive him from his driveway in another city 3+ hrs away to my driveway with no intervention. I feel like everyone’s opinion on how self-driving is going is still rooted in 2018 or something and no one has updated.

>I had a friend who had his Tesla drive him from his driveway in another city 3+ hrs away to my driveway with no intervention. I had anecdata that was data, and it said that full-self-driving is wishful thinking. We cool now?

Good luck on your journey. I think the world is going to surprise you, and you’d be better for opening your eyes a little wider.

Re: The next chapter of the Microsoft–OpenAI partnership

#256

Post by dreamcompiler »

IlikeKitties wrote: post #142
ramses0 wrote: post #126

Earlier quoted context omitted.

Jesus, we've gone from Eliza and Bayes Spam Filters to being able to hold an "intelligent" conversation with a bot that can write code like: "make me a sandwich" => "ok, making sandwich.py, adding test, keeping track of a todo list, validating tests, etc..." We might not _quite_ be at the era of "I'm sorry I can't let you do that Dave...", but on the spectrum, and from the perspective of a lay-person, we're waaaaay c…

I think "we" have accidentally cracked language from a computational perspective. The embedding of knowledge is incidental and we're far away from anything that "Generally Intelligent", let alone Advanced in that. LLMs do tend to make documented knowledge very searchable which is nice. But if you use these models everyday to do work of some kind that becomes pretty obvious that they aren't nearly as intelligent as th…

Yes, exactly. LLMs are lossy compressors of human language in much the same way JPEG is a lossy compressor of images. The difference is that the bits that JPEG throws away were manually designed by our understanding of the human visual cortex, while LLMs figured out the lossy bits automatically because we don't know enough about the human language processing chain to design that manually.

LLMs are useful but that doesn't make them intelligent.

Re: The next chapter of the Microsoft–OpenAI partnership

#257

Post by shon »

philipwhiuk wrote: post #79

> Once AGI is declared by OpenAI, that declaration will now be verified by an independent expert panel. > Microsoft’s IP rights for both models and products are extended through 2032 and now includes models post-AGI, with appropriate safety guardrails. Does anyone really think we are close to AGI? I mean honestly?

I honestly think that if you were to show the tech we have today, to someone at OpenAI back in 2015, they would say “we did it!!”

Outside of robotics / embodied AI, SOTA models have already achieved Sci-Fi level capability.

Re: The next chapter of the Microsoft–OpenAI partnership

#258

Post by crazygringo »

fadedsignal wrote: post #154
rapind wrote: post #148

Earlier quoted context omitted.

> Most people didn't think we were anywhere close to LLM's five years ago. That's very ambiguous. "Most people" don't know most things. If we're talking about people that have been working in the industry though, my understanding is that the concept of our modern day LLMs aren't magical at all. In fact, the idea has been around for quite a while. The breakthroughs in processing power and networking (data) were the ho…

I 100% agree with this. I suggest the other guy to check history of NLP.

Not sure what history you're suggesting I check? I've been following NLP for decades. Sure, neural nets have been around for many decades. Deep learning in this century. But the explosive success of what LLM's can do now came as a huge surprise. Transformers date to just 2017, and the idea that they would be this successful just with throwing gargantuan amounts of data and processing at them -- this was not a common viewpoint. So I stand by the main point of my original comment, except I did just now edit it to say 10 years ago rather than 5... the point is, it really did seem to come out of nowhere.

Re: The next chapter of the Microsoft–OpenAI partnership

#259

Post by enricotal »

“OpenAI is now able to release open-weight models that meet requisite capability criteria.”

Was Microsoft the blocker before? prior agreements clearly made true open-weights awkward-to-impossible without Microsoft’s sign-off. Microsoft had (a) an exclusive license to GPT-3’s underlying tech back in 2020 (i.e., access to the model/code beyond the public API), and (b) later, broad IP rights + API exclusivity on OpenAI models. If you’re contractually giving one partner IP rights and API exclusivity, shipping weights openly would undercut those rights. Today’s language looks like a carve-out to permit some open-weight releases as long as they’re below certain capability thresholds.

A few other notable tweaks in the new deal that help explain the change:

- AGI claims get verified by an independent panel (not just OpenAI declaring it).

- Microsoft keeps model/product IP rights through 2032, but OpenAI can now jointly develop with third parties, serve some things off non-Azure clouds, and—critically—release certain open-weights.

Those are all signs of loosened exclusivity.

My read: previously, the partnership structure (not just “Microsoft saying no”) effectively precluded open-weight releases; the updated agreement explicitly allows them within safety/capability guardrails.

Expect any “open-weight” drops to be intentionally scoped—useful, but a notch below their frontier closed models.

Re: The next chapter of the Microsoft–OpenAI partnership

#260

Post by dymk »

evtothedev wrote: post #18

> Microsoft’s IP rights now exclude OpenAI’s consumer hardware. While not unexpected, this is exciting and intriguing. And of course, looking forward to Microsoft's Zune AI.

Maybe we’ll get a wearable pin, look how well those have done so far
Post reply on HN