Former OpenAI employee quit to avoid 'working for the Titanic of AI'
cointelegraph.com
cointelegraph.com
One single person, from an appendix team, quit and speaks up about being unhappy with the direction of the company.
This happens all the time everywhere. The only mild relevance is that it's about today's Hot Topic company.
There aren't lessons to learn, besides the fact that we're wasting our time. The only risk around AGI is that people will trust it as much as they trust current AI, and then risk their own annihilation when they casually accept a blatantly hallucinated answer. It's like if HAL-9000 came preinstalled with Alzheimer's, and we're still adamant that we can trust him to operate the airlock. It's time to get real and admit that any risk imposed by AGI (or regular AI, for that matter) is a failure in implementation. The concept of alignment is a bald-faced marketing ploy intent on selling "AI safety" to people that don't know what AI even is.
1) alignment as portrayed by OpenAI along with the perception of intelligence from the current SoTA LLMs
2) alignment as a field of study to prevent catastrophe
I'm speaking about the latter and I don't think labeling alignment as a total facade is conducive to starting well on the path of developing guiderails for AGI. Thank the universe that it seems LLMs cannot scale to emergent intellect, because that lets us do a dry run. In my opinion, we're currently tanking the current go-around as legislation is ever farther behind while we watch an absolute arms race between nations and corporations.
Splash a dash of paint on the same moral and ethical questions debated ad-nauseam for very little money, sprinkle lies about how dangerous and advanced a de rigueur technique (LLM) which masquerades as a whole field (artificial intelligence) is, and by golly some silly Silicon Valley pump and dump will give you an office to pontificate in.
That doesn't detract from the utility of what is there already. I'm currently using it quite a lot as a "tutor" (of sorts) as I deep-dive into an area I have limited experience with. Like all tutors it gets some things wrong (which in turn reflects human misunderstandings). It is sufficiently good though at directing the flow, which in turn gives me Google terms to search.
I say this not as a rabid AI fan, I've come to it reluctantly, but its proving useful.
So I don't think shareholders will revolt (which means what? That the share price will drop?) - I think they'll just continue to provide (improved) utility and their user base will grow. This isn't like fusion where there's a binary measure of success.
Of course, like any business, there will be challenges along the way. Some survive those (Apple, MS, Google etc), some don't (Sun, DEC, Amiga). Predicting which ones succeed and which ones fail is hard. And seldom correlated to culture.
The article is fluff because it makes no conclusion, adds no data, is just a single ex-employees opinion. Meh.
What we are saying is that GenAI will never be able to go beyond code helper or assistant and the people hyping these tools continually promise AGI or something that can solve complex problems.
The ROI is just not there and eventually there will be a margin call.
There us plenty of value yo be unlocked well short of AGI.
Are investors excited? Yes. Is that excitement being fueled by hype? yes. Will excitement cool eventually? Sure. Will some investors, investing in the wrong thing, lose money? Sure.
But "revolt" (whatever that means) - no, IMO not.
The promises will be enough to get to usable AI as a helpful tool which makes money, without the perfectionism of something impossible.
In my analogy, personally I don't care about FSD, what you get is helpful. Dangerous is as overblown as EV fires. (think of human drivers and gasoline fires)