HNHacker News
TopNewBestAskShowJobs

skepticATX

1,464 karma · joined April 5, 2023

Software dev interested in security, distributed computing, and AI.
submissionscomments
skepticATX··on OpenAI Dissolves High-Profile Safety Team After Chief Scientist Sutskever's Exit
Completely agree. I think the AGI fantasy that took the world by storm a year and a half ago has died out.

Of course, there are still true believers in these companies. They were true believers years and decades ago. But everyone who was suddenly introduced to the concept is no longer interested as the frictions of real world progress become apparent.

skepticATX··on OpenAI Dissolves High-Profile Safety Team After Chief Scientist Sutskever's Exit
It really seems like executives have internally become disillusioned with the AGI fantasy. They’ll keep selling it, of course (although I have noticed quite the drop off in this type of language recently), but they don’t actually believe it any longer.
skepticATX··on Ilya Sutskever to leave OpenAI
Just the opinion of an outsider, so not worth very much. But Ilya seemed to be one of the few who actually believed in the mission. I’m sure it was hard for him to watch the company become so product focused.

OpenAI under Sam strikes me as completely disingenuous - and the constant hyperbolic tweeting by many OpenAI employees just reinforces that.

Too bad. While I don’t really think that OpenAI is on the right track for general intelligence, it certainly could have been a positive for the world.

skepticATX··on Veo
OpenAI hardly released gpt-4o. The demo yesterday was clearly a rushed response to I/O. It’s quite possible that Google will ship multi-modality features faster than OpenAI will.
skepticATX··on GPT-4o
All of that is awesome, and makes for a better product. But it’s also primarily an engineering effort. What matters here is an increase in intelligence. And we’re not seeing that aside from very minor capability increases.

We’ll see if they have another flagship model ready to launch. I seriously doubt it. I suspect that this was supposed to be called GPT-5, or at the very least GPT-4.5, but they can’t meet expectations so they can’t use those names.

skepticATX··on GPT-4o
The increased expectations are a direct result of LLM proponents continually hyping exponential capabilities increase.
skepticATX··on GPT-4o
Very impressive demo, but not really a step change in my opinion. The hype from OpenAI employees was on another level, way more than was warranted in my opinion.

Ultimately, the promise of LLM proponents is that these models will get exponentially smarter - this hasn’t born out yet. So from that perspective, this was a disappointing release.

If anything, this feels like a rushed release to match what Google will be demoing tomorrow.

skepticATX··on Claude 3 beats GPT-4 on Aider's code editing benchmark
No. Humans can do System 2 thinking.

Quite frankly this view ignores decades of cognitive science that clearly demonstrates that the brain is not just predicting next tokens.

skepticATX··on The New Inflection
Mustafa wasn’t exactly on his best behavior at DeepMind, if the allegations are to be believed. It really is surprising that Microsoft would hire him.
skepticATX··on Microsoft hires DeepMind co-founder Suleyman
It would seem that Microsoft doesn’t care about the history of abuse allegations at DeepMind.
skepticATX··on Claude 3 model family
The results really aren’t striking enough that it’s clear that this model blows GPT-4 away. It seems roughly equivalent, give or take a bit.

Why can we still not easily surpass a (relatively) ancient model?

skepticATX··on "AI will cure cancer" misunderstands both AI and medicine
> humanity, given 1000 years

The problem is that this makes almost anything sound reasonable, besides things that violate the laws of physics.

I don't think that thinking this far out is anything but pure speculation. Whether the mechanism is human intelligence, AI, or magic.

skepticATX··on "AI will cure cancer" misunderstands both AI and medicine
The predominate view of AI is that it is essentially magic. Which of course couldn’t be further from the truth.

But I find these arguments much easier to understand if you look at it through that lens. Of course magic can cure cancer because… it’s magic!

skepticATX··on Top AIs still fail IQ tests
Right. And the industry has slowly realized this. The "AGI" hype has died down since reaching a fever pitch when GPT-4 was released. Result after result has shown that LLMs have extreme difficulty generalizing beyond what they've been trained on. Which really isn't much of a surprise, if you think about how deep learning works.

Of course, the major labs and AI influencers are still publicly shouting from the rooftops, because that is what sells. But it's getting harder to deny the obvious.

Ultimately, the problem is that there is really no consequence for being wrong, but tremendous upside for this kind of magical thinking. Someone should start tracking predictions so we can ignore those who continually get things wrong.

skepticATX··on Tyler Perry Puts $800M Studio Expansion on Hold After Seeing OpenAI's Sora
Seems like quite the overreaction to a research demo.

Yes, it will get better. But generating anything close to coherent film / television show is a tall task.

For all of the progress in image generation we still seem somewhat stagnant when it comes to fixing hallucinations. Video generation is a much harder problem than image generation, so it's not super clear to me we're very close to this having a material impact on Hollywood.

skepticATX··on What Was Sora Trained On? Creatives Demand Answers
We don't have to treat humans and machines the same.

Even if it were true that these models learn exactly like humans (which is patently false), we can still differentiate between the two.

Society is built by and for humans. If we want to empower humans over machines, that seems perfectly acceptable.

skepticATX··on What Was Sora Trained On? Creatives Demand Answers
I don't know if it's illegal, because I'm not a lawyer.

I do know that it is completely unethical. A few large companies should not be able to package up and sell the totality of humanity's creative output at the expense of humanity itself.

The backlash is certainly just beginning. Thus far, AI usage beyond brief exploration have mostly been driven by tech folks. As more people outside of the bubble get clued into what is happening, I think that reputations will take a serious hit.

skepticATX··on AI hiring tools may be filtering out the best job applicants
Quite frankly, using generative AI to screen candidates should be illegal until/unless the technology becomes more steerable and explainable.

It's not a question of if these technologies are biased, it's only a question of how biased they are.

skepticATX··on I'm an Old Fart and AI Makes Me Sad
> it won’t be made by typing your idea into DALL-E, it’ll happen with complex workflows on fine-tuned models.

I'm not convinced that this is what will happen long-term, but if it is, I'd completely agree that it's a new art form.

skepticATX··on I'm an Old Fart and AI Makes Me Sad
Photography still takes a ton of skill though. It's still ultimately a human creation. AI art is the antithesis of this.
skepticATX··on I'm an Old Fart and AI Makes Me Sad
Agreed. I think that the concern about disinformation stemming from AI images/video misses the bigger picture.

Humanity will adapt to these new technologies. We'll begin to trust what we see less and less. Which will prevent us from being fooled, but we also lose a massive part of the human experience in the process.

Soon the days of being able to view a photograph, a video, artwork, and appreciate it as human ability at its pinnacle will be gone.

skepticATX··on I'm an Old Fart and AI Makes Me Sad
Many of the points made in this article apply to generative AI specifically. Which makes sense, because it's the flavor of the decade. But I think it's worth pointing that out, because AI does not inherently have to be inaccessible and unexplainable.

Beyond that, it is up to us to guide AI development and deployment. If we use it to crush the human spirit (and it sure seems like we're hell-bent on doing that right now), that's more of an indictment on AI leaders, and in a broader sense all of humanity, than it is on the technology itself. Nothing is inevitable, despite what some in the industry want to have you believe.

skepticATX··on Sora: Creating video from text
This is what lots of folks said about image generation. Which is now in many ways “solved”. And society has easily adapted to it. The same will happen with video generation.

The reality is that people are a lot more resourceful / smarter than a lot of us think. And the ones who aren’t have been fooled long before this tech came around.

skepticATX··on Hi everyone yes, I left OpenAI yesterday
Frankly, OpenAI seems to be losing its luster, and fast.

Plugins were a failure. GPTs are a little better, but I still don't see the product market fit. GPT-4 is still king, but not by that much any more. It's not even clear that they're doing great research, because they don't publish.

GPT-5 has to be incredibly good at this point, and I'm not sure that it will be.

skepticATX··on Americans don't believe middle class can afford homes
In most ways the lives of middle class Americans are better than that of their counterparts 30 years. We have an idealized view of what prior generations’ middle class looked like.

Home ownership is the one exception. We need to build denser, and we need to build actual communities that are livable without needing to drive 30 minutes to go to the grocery store.

The middle class of today has every right to be infuriated at our housing inaction.

skepticATX··on Yann LeCun: Human-level artificial intelligence is going to take a long time
Yann has been by far the most clear-headed of all the AI leaders since the ChatGPT hype started.

I’m not a huge fan of Meta but it’s hard not to like the work they are doing in AI. High expectations for their future as long as Yann is around.

skepticATX··on Mark Zuckerberg’s new goal is creating artificial general intelligence
I don’t think that’s Meta’s viewpoint, considering FAIR is run by Yann LeCun who has been quite vocal about the limitations of what we currently have.
skepticATX··on Bing Gained Less Than 1% Market Share Since Adding Bing Chat
The “Google is going downhill” narrative only exists in the tech bubble. No one actually thinks this way in the real world.

Generative AI features can enhance search somewhat but aren’t a replacement. It’s not enough of a driver on its own to take market share from Google.

skepticATX··on AlphaGeometry: An Olympiad-level AI system for geometry
Every single human has abstractions that are unique to them. Your world model isn’t the same as mine.

It’s just that usually these abstractions are fuzzy and hard to formalize, so they aren’t shared. It doesn’t mean that they don’t exist.

skepticATX··on A developer's second brain: Reducing complexity through partnership with AI
This is a surprisingly balanced article, which I no longer expect to come out of Microsoft.

I suspect that companies focusing on augmentation instead of replacement will be much more successful in the long run. Good to see that GitHub at least seems to get that.

← PreviousPage 3 of 9Next →