It sucks that I would love to be excited about this... but I mostly feel anxiety and sadness.
While there are those that are excited, the world is not prepared for the level of distress this could put on the average person without critical changes at a monumental level.
That would bug me, if I were you.
You can't bullshit your way through this particular benchmark. Try it.
And yes, they're wrong. The latest/greatest models "make shit up" perhaps 5-10% as frequently as were seeing just a couple of years ago. Only someone who has deliberately decided to stop paying attention could possibly argue otherwise.
I have noticed it's great in the hands of marketers and scammers, however. Real good at those "jobs", so I see why the cryptobros have now moved onto hailing LLMs as the next coming of jesus.
I do find, however, that the newer the model the fewer elementary mistakes it makes, and the better it is at figuring out what I really want. The process of getting the right answer or the working function continues to become less frustrating over time, although not always monotonically so.
o1-pro is expensive and slow, for instance, but its performance on tasks that require step-by-step reasoning is just astonishing. As long as things keep moving in that direction I'm not going to complain (much).
/s
"It is ceasing to be a matter of how we think about technics, if only because technics is increasingly thinking about itself. It might still be a few decades before artificial intelligences surpass the horizon of biological ones, but it is utterly superstitious to imagine that the human dominion of terrestrial culture is still marked out in centuries, let alone in some metaphysical perpetuity. The high road to thinking no longer passes through a deepening of human cognition, but rather through a becoming inhuman of cognition, a migration of cognition out into the emerging planetary technosentience reservoir, into 'dehumanized landscapes ... emptied spaces' where human culture will be dissolved. Just as the capitalist urbanization of labour abstracted it in a parallel escalation with technical machines, so will intelligence be transplanted into the purring data zones of new software worlds in order to be abstracted from an increasingly obsolescent anthropoid particularity, and thus to venture beyond modernity. Human brains are to thinking what mediaeval villages were to engineering: antechambers to experimentation, cramped and parochial places to be.
[...]
Life is being phased-out into something new, and if we think this can be stopped we are even more stupid than we seem." [0]
Land is being ostracized for some of his provocations, but it seems pretty clear by now that we are in the Landian Accelerationism timeline. Engaging with his thought is crucial to understanding what is happening with AI, and what is still largely unseen, such as the autonomization of capital.
Sure, there will be growing pains, friction, etc. Who cares? There always is with world-changing tech. Always.
That's right. Who cares about pains of others and why they even should are absolutely words to live by.
What you are likely doing, though, is making many more future humans pay a cost in suffering. Every day we delay longevity escape velocity is another 150k people dead.
I'd rather just... not die. Not unless I want to. Same for my loved ones. That's far more important than "wealth inequality."
Senescence is an adaptation.
"You should die because cities will get crowded" is a less terrible argument but still a bad one. We have room for at least double our population on this planet, couples choosing longevity can be required to have <=1 children until there is room for more, we will eventually colonize other planets, etc.
All this is implying that consciousness will continue to take up a meaningful amount of physical space. Not dying in the long term implies gradual replacement and transfer to a virtual medium at some point.
If you take this as an axiom, it will always be true ;).
“Oh well, I guess I can’t give the opportunities to my kid that I wanted, but at least humanity is growing rapidly!”
Everyone has always worried about this for every major technology throughout history
IMO AGI will dramatically increase comfort levels, lower your chance of dying, death, disease, etc.
People aren’t really on uproar yet, because implementations haven’t affected the job market of the masses. Afterwards? Tume will show.
Like in general I totally agree with you, but I also understand why a person would care about their loved ones and themselves first.
Beyond immediate increase in inequality, which I agree could be worth it in the long run if this was the only problem, we're playing a dangerous game.
The smartest and most capable species on the planet that dominates it for exactly this reason, is creating something even smarter and more capable than itself in the hope it'd help make its life easier.
Hmm.
>But while the “making AGI” part of the mission seems well on track, it feels like I (and others) have gradually realized how much harder it is to contribute in a robustly positive way to the “succeeding” part of the mission, especially when it comes to preventing existential risks to humanity.
Almost every single one of the people OpenAI had hired to work on AI safety have left the firm with similar messages. Perhaps you should at least consider the thinking of experts?
You and I will likely not live to see much of anything past AGI.
The people experiencing the growing pains, friction, etc.
For one, I found AI coding to work best in a small team, where there is an understanding of what to build and how to build it, usually in close feedback loop with the designers / users. Throw the usual managerial company corporate nonsense on top and it doesn't really matter if you can instacreate a piece of software, if nobody cares for that piece of software and it's just there to put a checkmark on the Q3 OKR reports.
Furthermore, there is a lot of software to be built out there, for people who can't afford it yet. A custom POS system for the local baker so that they don't have to interact with a computer. A game where squids eat algae for my nephews at christmas. A custom photo layout software for my dad who despairs at indesign. A plant watering system for my friend. A local government information website for older citizens. Not only can these be built at a fraction of the cost they were before, but they can be built in a manner where the people using the software are directly involved in creating it. Maybe they can get a 80% hacked version together if they are technically enclined. I can add the proper database backend and deployment infrastructure. Or I can sit with them and iterate on the app as we are talking. It is also almost free to create great documentation, in fact, LLM development is most productive when you turn up software engineering best practices up to 11.
Furthermore, I found these tools incredible for actively furthering my own fundamental understanding of computer science and programming. I can now skip the stuff I don't care to learn (is it foobarBla(func, id) or foobar_bla(id, func)) and put the effort where I actually get a long-lived return. I have become really ambitious with the things I can tackle now, learning about all kinds of algorithms and operating system patterns and chemistry and physics etc... I can also create documents to help me with my learning.
Local models are now entering the phase where they are getting to be really useful, definitely > gpt3.5 which I was able to use very productively already at the time.
Writing (creating? manifesting? I don't really have a good word for what I do these days) software that makes me and real humans around me happy is extremely fulfilling, and has allevitated most of my angst around the technology.