Hacker News Reader

Hacker News

Comment by rideontime | Hacker News Reader

Full thread

rideontime

Sam Altman says a lot of things.

View on HN

Ok, but please don't post unsubstantive comments to Hacker News. It's tedious and evokes worse from others.

He's been good and bad for this field.

Good:

- Attracted a lot of attention, investment, and engineering to this field. It's unlikely we'd be where we are today without OpenAI shocking the world with its demos.

- Built some salient products like ChatGPT

Bad:

- Tried to get congress to create a regulatory moat for AI (perhaps when he realized there was no moat for himself)

- Tried to dry up funding for other startups (again to try to create a moat)

- Over-hypes and over-promises what his team has done and what can be delivered (a boon to his fundraising, but making funding frothy and disconnected from fundamentals)

- Engaged in very public-facing drama around his board, his company's nonprofit status, etc.

- Telling congress he has no equity in OpenAI and is doing everything with no reward, yet driving around in a $5M Koenigsegg Regera supercar and getting a generous comp package put together for the new corporate governance structure.

- Creepy stuff like eyeball-scanning Worldcoin is pitched as essential in a post-AGI world, but transacting on non-revocable biomarkers seems like some 90's televangelist's sermon about the "Mark of the Beast"

> Tried to get congress to create a regulatory moat for AI

He was fundraising in the Congress. The whole “stop me before I shoot grandma” shtick was stupid and aimed at Silicon Valley. The moment states and countries started regulating AI he was caught flat footed.

I'm wondering how much of my repulsion to AI is Sam Altman himself. Recently I've been trying to be more open-minded with using LLMs in my workflows and have found some minor performance gains here and there. However the hype around AI is frankly embarrassing.

I think this extends to a lot of normal people’s instinctive hatred of AI: they dislike it in part because they perceive it as being pushed on them by some pretty slimy people

wait until it crosses a few thresholds we’re fast approaching. ai that performs at intern level vs phd vs domain leading expert level. ai that has the full context of what youre doing immediately accessible. ai that can agenticly autonomously navigate the computer environment without stumbling blocks.

i think the difference between intern level ai and domain leading expert ai is a few algorithmic adjustments away using a type of reasoning reinforcement framework (like GRPO) which deal with competing signals in a better way. instead of averaging them, it reasons which contextually signal should take precedent. its the difference between lets say taking a vote among the common populace on how to build a nuclear power plant and finding an expert and figuring out where the experts decisions should take precedent and where other experts decisions should override.

square that away and the embarrassing feeling on the promise of ai should wash away.

> ai that performs at intern level vs phd vs domain leading expert level.

These are different things:

- Regurgitating advanced text that has been shifted into a shape matching your query

- Understanding intimately the $100M screw to turn

obviously theres also a multimodality gap to be overcome to intimately understanding the $100M screw to turn, but i suspect most reasoning that matters has already been translated and is embedded into/in words. i wouldn’t underestimate the amount of useful knowledge that exists in embedding advance texts into an LLm model. the challenge is contextually hierarchalizing it (a matter of reasoning) and decoding it back into reality (words are dimensional squished encodings of reality).

How do you know that these are different things? They could be, I genuinely don’t know, but I’m not sure where people are getting these kind of confident assertions about what modern architectures could never do. Would you have predicted in 2020 that photorealistic text to image generation was within the scope of current theory?

The mathematical definition of a language model is the probability distribution of tokens that follows the previous context. It's literally deciding the most probable response, which while at many times may match the correct response, is not a 1 to 1.

welcome to earth where 98% of forum/political interactions are confident assertions from nowhere used to dismiss people. enjoy your stay ;).

> Would you have predicted in 2020 that photorealistic text to image generation was within the scope of current theory?

Yes, and I've been working in this area with excitement since about that time.

The physics of optics are well understood. We've been writing ray tracers for forever and coming up with clever hacks like Blinn–Phong, PBR, etc. for ages. SIGGRAPH has always felt like tangible magic. We have had the map in our hands and now we're coming up with new ways to traverse a familiar landscape.

Reasoning is an undiscovered country. There are lots of exciting claims being made, but nothing concrete.

I expect lots of advancements in signal processing, spatial computing, and beyond because those things are obvious and intuitive.

Yeah. I gave Claude a spin today; it helped me to make a tech spec I'd written a little more concise and professional. But doing so required carefully reading and tweaking the output to maintain accuracy. I don't know if it sped anything up, but it reduced the cognitive effort a bit.

Big part of my annoyance is the term "AI" itself which you can say to mean anything, everything and nothing. It's something that's used to oversell/hype it to grab money.. which is fine. But if engineers like us can calm the crazy rhetoric down to "llms" or "text completion" "Image-gen api" that's already a leg up in thinking clearly. Like the question "is ai going to put people out of jobs" -> "is this new crazy good text completion model going to put people out of jobs" already gets us out of the weeds somewhat.

Yeah its definition has been diluted to the point of useless now.

> Big part of my annoyance is the term "AI" itself which you can say to mean anything, everything and nothing.

Pretty sure that's the point now.

You're selling a "solution" not a "product". So you don't want to market AI as some nice robot that can move car doors around. You want to market AI as a vibe that fixes enterprise problems. So now all the companies that want your AI both have to pay for AI and also your time to actually build a product for their problem.

Would be happy to go through my life hearing no more from him.

The forest is cold, but peaceful.

Also not being reminded that technology indistinguishable from magic will happen in the next 5 years, or whatever carrot-in-a-stick money grab.

I say a lot of things, Monica. I say a lot of things.

https://youtu.be/XM7_eqtljUg?si=d27SGVUiZ7oZJCA3

At least he doesn't do a lot of "gestures".

The exact post I was about to make, glad to see it at the top, lol.

I keep reading people on the internet mistrusting him with a lot of confidence, but I haven't heard of any tangible evidence that he's lying about anything.

Can you name a couple of examples of the things he said that we know are lies? Or is it all just people making uninformed assumptions or being snarky?

Changing his company from a non-profit to a for-profit once he saw the $$$, seems pretty untrustworthy to me.

OpenAI is not changing to a for-profit. It[1] always was a for-profit entity, owned by a non-profit.

The big change is that the non-profit no longer owns the entirety (or even a majority) of the for-profit entity; it is now a minority owner.

[1] OpenAI as we know it today. OpenAI was once just the non-profit entity, but back then was just an AI think tank. In 2019, it formed the for-profit corporation as a subsidiary to raise money and build the tech that we now know about (and make money from any products built on that tech).

I don't know if this is technically a lie, but he said that he'll purchase electricity made from fusion power within the next couple of years.

I don't know if he believes that himself. But I can tell you with extreme high confidence that this is not going to happen, and it's not even close to anything remotely realistic.

Secretly funding the FoundationMath benchmark, contributors unaware of the COI, having access to the questions and answers with a "verbal agreement" not to train on it.

https://techcrunch.com/2025/01/19/ai-benchmarking-organizati...

The main thing is OpenAI itself. Altman has long pitched it as an open, non-profit and raised hundreds of millions on that. Turns out it's not open and now has been positioned as a for-profit entity owned by a non-profit that is trying to convert itself to a for-profit.

On top of that he repeatedly used the fact that he had no equity to push back on criticism and to appear more altruistic and trustworthy. Turns out, that was just part of the con.

The literal bait-and-switch of "Oh hey Open AI is nonprofit I promise guys!" and then turning around and switching it to profit.

That alone is enough to walk away from him.

One doesn't have to be a liar to be untrustworthy

Why was he removed from leadership positions at Loopt, YC, and OpenAI?

https://www.disconnect.blog/p/kara-swishers-story-about-sam-...

Sam Altman is currently being sued by his sister for raping and sexually abusing her for 9 years. He is denying this, I guess that counts as lies.

https://www.cnn.com/2025/01/08/business/sam-altman-denies-si...

He says a lot of things that are untrue... but this... its true. Its why Deepthink makes no diffrence.

Reply on news.ycombinator.com