The best outcome right now is for OpenAI to open source GPT-4 architecture and its training data, then release the weights. Would at least take LLM dominance away from Microsoft.
But seeing the shenanigans of the past few days and how MSFT failed to secure a board seat for their $$, it would not surprise me.
A man can dream...
As far as I'm concerned literally everyone else is "one of the good guys" when the other side we're talking about is M$ and the psychopaths that are involved with it, regardless of their actions.
the new CEO wanted to focus on something else than LLMs, because it's believed to be a dead end, and allocate more time and resources into grounding AI.
they don't want AI to slow down, they want to try out something else and not put all their eggs in one basket. extremely big difference
https://x.com/eshear/status/1703178063306203397?s=46&t=yWOhq...
I see nothing in that thread that says what you’re talking about.
OpenAI's best asset isn't GPT-4. GPT-4 is just an interim step to GPT-5 and 6 and 7 and so on.
Imagine we had this conversation when GPT-3 was out. Just open-source it and disband the team! And what? Nothing.
The people are always the greatest asset of a software company. Always. That's even the case with AI. At least... until I guess we get an AGI which is so powerful that it can self-improve, then I guess screw the people. But then that applies not just to the OpenAI team, but the world as a whole.
For the HN readers commenting on this.
People think you open-source something and it starts improving rapidly. But that's not exactly how it works. The first push is almost always commercial. And only when the product is mature, well understood in and out, and people get bored of paying for it, then open-source initiatives start showing up to democratize this product type for everyone without dependency on a single vendor.
Think like Unix -> Linux. Linux is largely stagnant, but that's also because it's "done". Sure tons of patches flow in, but if you need a major architectural change, with breaking changes, FOSS can't deliver that most of the time, if ever.
Are we done developing AI? Nothing major to add, that's it? Hell, no. We're merely starting. And so... we need to see what commercial companies (even under the guide of not-for-profit ones) can do, before we start demanding open-source versions.
You mean destroy the rubble that's currently left in the offices of the company formerly known as OpenAI? If 500-750 key employees have departed to a competitor, what's left to protect?
Yeah, it would 80% be a "fuck you" move, but the other 20% would represent a pivot to an "open source" model like the name once implied
Or they would be triggering a billion-dollar lawsuit that would surely bankrupt the company.
https://news.ycombinator.com/item?id=38349969
Even if there isn't anything explicit, there may be an implicit promise that a strategic partner won't share all the company secrets with the partner's competitors.
See also https://twitter.com/karaswisher/status/1726598360277356775 which was merged into that discussion above
I'm deeply concerned about it falling into the arms of private actors, specially the worst of them, M$.
They have their own kind of AI technology which serves different purposes and we probably know nothing or very little about it. It's not like they will be sharing their tech with the World, much less open sourcing it.
Besides, state actors probably already have tech that far surpasses GPT and they're probably using it for a lot more than making people feel clever for copy/pasting a CRUD app together for their useless startup.
What do you mean by "moral"? State actors have interests, major corporations also have their interests. Very few of them can be called moral.
Big tech was always about power and money. No morals there.
> and its training data
I'm not sure you can opensource something you don't actually hold any copyright to..
While undermining its safety argument. Better: open it to partners, subject to vetting and contractual terms regarding its use, in exchange for cash and/or compute.
For someone not knowledgeable in this, why is this the case?
So you open this to the world - what's the stop the Chinese government from weaponizing it to both suppress their own citizens and attack foreign adversaries?
Except... it does and the net effect was seen just this week as Xi started softening his tone realizing that his economy cannot keep up without access to advanced technology.
In other words, for many people in the world that wouldn't be an issue, it's already in the hands of a country that had invaded or intervened far more countries than China in the past few decades
I would care what China would do if China was going around the world killing millions of people in the name of "national security".
Alas, so far that privilege went to Great Britain, France (Napoleon), Germany, US and recently Russia and Isreal.
Still no China on that radar.
That belief underlies the OpenAI's prevailing nonprofit mission and is the reason for its existence.
Prematurely loosing AI amongst the lay people is their nightmare scenario. Only the technocrat elect can perform the ritual safely and only then if they work together and avoid earthly temptation for power and money.
Currently the for-profit models are Nerfed to not allow bigoted or sexist (against women) text to be generated but that seems to be it. What kind of text are we being protected from?
Sexist etc content is a footnote compared to the potential for an automated, proven, and tuneable technology of mass persuasion and impersonation.
So - massively addictive adtech. Reliable and effective political and ideological propaganda. Bot farms and "credible" fake media content produced at scale.
Imagine those being used against a relatively defenceless population to creative division, undermine personal and collective confidence, and encourage stochastic terrorism and other kinds of violence.
The risk with AI is monopolization by humans using it for nefarious purposes. This risk is reduced through openness and transparency.
Which is just political marketing for regulatory barriers to entry lobbying.
But neither seem likely to want to open anything meaningful that is is AGI related-- locking the public out of powerful AI is a point in which all the major 'safety' factions are aligned. (well not the monopolization concern faction, but they're not a player here and usually don't hide behind the safety banner)
You're saying that would be the best outcome, but... why?
- The OpenAI staff stays and the current board is replaced by Microsoft
- The OpenAI staff leaves and goes work for Microsoft.
Either way Microsoft has 100% control over this technology, thus the best possible outcome is OpenAI releasing it.