Sama is smarter, yes, but both are egotistical narcissists for whom winning is more important than the cause. Both are behaving in wildly unethical ways. Both try to manipulate policy to whitewash their behavior in the eyes of the public. 🤷
I have a completely unrelated question: I've noticed that lately you're using GPT-4o for your featured images. I recall there was a period where you were consciously moving away from AI images towards the public domain. What ended up changing your mind?
(No judgment here whatsoever, pure curiosity. I've been using AI-generated images for my featured images since the start.)
Good question! I actually answered it in the post before this one but long story short I think it no longer makes any sense to resist AI writing so why resist AI images. I won't do it myself because I like writing but I won't look down on people using AI to the degree they consider (I'd like to avoid AI writing, just because it's harder to assess the quality as a reader)
Oh damn, I must've missed that post - gotta catch up on my reading!
My stance on this has always been that AI is especially well-suited for filling in gaps in one's skillset. I see myself as a writer but definitely not a designer/illustrator, so I'm happy to use AI for that aspect.
Someone else might have a keen visual eye but struggle to get their message across in written form, so they'll use chatbots for that.
Of course, this doesn't take into account any iffy ethical issues in terms of unlicensed training data, copyright material, etc. The cautious optimist in me would like to think that we'll figure this out from a legal and societal standpoint though.
Yep, that's it. I also defend the idea that we should separate individual stances from collective fights. I don't think it's hypocrisy to use ChatGPT and at the same time protest that AI companies' practices should not be illegal or at least unethical. I know it's a fine thread to navigate and it's always safer to just restraint oneself but I think that's merely virtue signaling: one can do a lot collectively even if individually exists in apparent contradiction and then another person could be very angry at AI and then do nothing of utility to change the status quo
1. One month ago you wrote: "Google Is Winning on Every AI Front: Neither OpenAI nor Anthropic have a chance at this point". Today you write: "Sam Altman Is Going to Win". Which is it? I'm pulling for Anthropic since they take transparency and safety far more seriously than anyone else, and because Claude is genuinely more helpful than the other models -- at least for programming.
2. Sam Altman is one of the most dishonest players in Silicon Valley, and I hope he doesn't win. He _is_ a very savvy operator though, and gives very helpful startup advice.
Yep, I'm drawing a line between winning at the technological level and at the "character" level. I find both those arguments compatible (I understand it can be confusing!)
It seemed AI-written rather than human-written, which I'd never thought before on any previous post. Maybe just me though, but I can't recall ever reading lines equivalent to many in here, aha.
Not necessarily the entire post. Just seemed a really different style from your typically more laid back/conversational-style posts.
Re your request, this is what I got:
"While the plebeian masses are distracted by the ephemeral glint of consumer-grade AI chatbots, the true battle for algorithmic supremacy is waged in the deep substrata of computational physics. Here, the titans of tensorflow grapple not with mere language models, but with the very architecture of synthetic cognition itself—a theatre where only those forged in the fires of foundational science dare tread. The future belongs not to the charming impresario or the fearful prophet, but to the quiet architect who understands that reality, digital or otherwise, bends only to the will of mathematical truth"
I do. There's parts in the above that I think are more OTT, and parts in the article I think are more OTT.
I'm not sure on it's value though - I'm only one reader. The main thing for me was just that it seemed a complete 180 from your usual style.
For example, I (personally) can't recall ever reading sentences like the below from your blog:
- "Their work echoed the same intellectual ambition and richness of spirit I might have felt toward Einstein’s relativity or the quantum mechanics of Schrödinger, Bohr, and Dirac"
- "He just keeps fleeing forward, chased by his own shadow—the shadow of a long string of mistakes, the latest of which was a disaster of Babelian proportions that, while negligible in consequence, left him exposed"
- The entire 'Nerve' segment
I wouldn't put too much weight in just my opinion though.
Believing Musk to be a positive force for the world after we've seen him cheer on AFG while doing Nazi salutes and getting Trump elected certainly clarifies what sort of world you are hoping for
Do you think that Sam Altman is more dangerous than someone who did a Nazi salute with perfect form at the inauguration?
Maybe you haven’t seen video of the salute and you’re just basing your opinion of the salute on articles that called it awkward hand waving that is being misinterpreted? Because it definitely 100% was a Nazi salute, no doubt whatsoever.
I’m not trying to do some gotcha thing here. I probably comment on your blog more than any single other reader lol. You’ve got my loyalty.
The exact phrase was “I think Musk is closer to being a positive force in the world”
I hear that closER, not literally a positive force, just closer to being one.
But you should know that if you’re saying that Musk is better for the world than Altman, that you’re suggesting that Altman is more dangerous than Nazi-ism.
That’s a pretty bold position to take. I could see how it might be true… AI could create a permenant dictatorship, p(doom) all that. I’d want to hear more about it from you, though.
Hey Geoffe, thanks for the question. Nazism is the secular version of the devil - except perhaps complete destruction of humanity I can't think of anything worse. Yes, I saw the vid. Elon Musk has grown severely unhinged and the salute was real and intentional. However, unhingedness may not be as problematic as master manipulation (i.e., signaling is not the same as doing). That's my reasoning at least! (That said, if Musk turns out to be a real hands-on nazi and aims to influence the political sphere on those terms, then that's a completely different scenario than what I'm assessing here.)
FWIW I don’t think Musk is a hands-on Nazi either. I think it was signaling too, and that his rationale was cynical, not ideological.
I think he did the salute to excite and titillate some of his worst allies and strike some serious, paralyzing fear into some of Trump’s enemies.
The whole first 100 days in office seemed like a propaganda assault more than anything else. Just a bunch of stuff that probably won’t hold up in court, done to throw people off balance and terrify them.
Their strategy seemed to rely on so much sacred cow desecration that when someone more liminal (eg the algorithmic bridge) says something fairly innocuous (eg “closer to positive”), it instantly creates a whole fucking issue that only serves Trump and Musk in the end. They desperately want attention and fear sent their way because it’s the root of their power and authority.
Anyway, way outside the scope of this blog.
Thanks for your explanation.
I had a feeling you’d get in trouble for that sentence. Hope this is the worst of it.
I'm used to this kind of trouble tbh hahaha. Anyway, glad you asked (not being American - ie, not having a special predilection to any particular side - is both a blessing and a curse when it comes to these topics)
signing the secular version of the devil to his allies seems to presuppose some terrible (and correct) things about his allies, allies that he financially, politically, and via propaganda supports.
I'd love to see where Altman is the most dishonest player in SV. Musk's recent political moves for me totally disqualify him as a potential "good guy" - his latest behaviour is pure evil and pushing for Trump when you built your persona on saving the planet is as moronic as it is ironic.
I actually think that Altman has this true 360 vision - based on his time spent at YC and his other investments that none of the others have. And he does things with a lot of cool, like he's having fun with his colleagues - just look at the demos he did with the teams. While building a company from scratch with 500 million users.
Like you I have lots of respect for Hassabis, who is the only true genius of the lot.
Not sure he's the most dishonest player though I just... Don't trust him. The "hunch" is not a rhetorical figure but quite literally how my intuition speaks about him
Missed opportunity not using Charisma, Uniqueness, Nerve, and Talent as the four characteristics.
Five letters each, I'm a man with a weakness for aesthetics (...oh wait!)
Oof!
In the Silicon Game of Thrones - Sam = Cersei Lannister, Elon = Daenerys Targaryen, Dario = Jon Snow, Mark = Stannis Baratheon
This is actually quite good, especially re Elon's turn from mostly good to mostly unhinged (Jon is too benevolent for anyone to be him though)
In this scenario doesn't Dario "win". Or where Bran fit in?
Sama is smarter, yes, but both are egotistical narcissists for whom winning is more important than the cause. Both are behaving in wildly unethical ways. Both try to manipulate policy to whitewash their behavior in the eyes of the public. 🤷
There is no honor in this game anymore. It’s sad.
I have a completely unrelated question: I've noticed that lately you're using GPT-4o for your featured images. I recall there was a period where you were consciously moving away from AI images towards the public domain. What ended up changing your mind?
(No judgment here whatsoever, pure curiosity. I've been using AI-generated images for my featured images since the start.)
Good question! I actually answered it in the post before this one but long story short I think it no longer makes any sense to resist AI writing so why resist AI images. I won't do it myself because I like writing but I won't look down on people using AI to the degree they consider (I'd like to avoid AI writing, just because it's harder to assess the quality as a reader)
Oh damn, I must've missed that post - gotta catch up on my reading!
My stance on this has always been that AI is especially well-suited for filling in gaps in one's skillset. I see myself as a writer but definitely not a designer/illustrator, so I'm happy to use AI for that aspect.
Someone else might have a keen visual eye but struggle to get their message across in written form, so they'll use chatbots for that.
Of course, this doesn't take into account any iffy ethical issues in terms of unlicensed training data, copyright material, etc. The cautious optimist in me would like to think that we'll figure this out from a legal and societal standpoint though.
Yep, that's it. I also defend the idea that we should separate individual stances from collective fights. I don't think it's hypocrisy to use ChatGPT and at the same time protest that AI companies' practices should not be illegal or at least unethical. I know it's a fine thread to navigate and it's always safer to just restraint oneself but I think that's merely virtue signaling: one can do a lot collectively even if individually exists in apparent contradiction and then another person could be very angry at AI and then do nothing of utility to change the status quo
Where there is smoke, there is fire and Altman smells of burnt wood.
1. One month ago you wrote: "Google Is Winning on Every AI Front: Neither OpenAI nor Anthropic have a chance at this point". Today you write: "Sam Altman Is Going to Win". Which is it? I'm pulling for Anthropic since they take transparency and safety far more seriously than anyone else, and because Claude is genuinely more helpful than the other models -- at least for programming.
2. Sam Altman is one of the most dishonest players in Silicon Valley, and I hope he doesn't win. He _is_ a very savvy operator though, and gives very helpful startup advice.
Yep, I'm drawing a line between winning at the technological level and at the "character" level. I find both those arguments compatible (I understand it can be confusing!)
He does play the game well but the "most dishonorable wins" is a very bad timeline.
Interesting post, but it came across very AI-y to me. Doesn't seem consistent with your typical style?
Interesting comment - can you clarify what do you mean with "AI-y"?
It seemed AI-written rather than human-written, which I'd never thought before on any previous post. Maybe just me though, but I can't recall ever reading lines equivalent to many in here, aha.
Do you really think this sounds AI written? Then I will ask you, please, to show us what you're capable to get from ChatGPT - I'm genuinely interested
Not necessarily the entire post. Just seemed a really different style from your typically more laid back/conversational-style posts.
Re your request, this is what I got:
"While the plebeian masses are distracted by the ephemeral glint of consumer-grade AI chatbots, the true battle for algorithmic supremacy is waged in the deep substrata of computational physics. Here, the titans of tensorflow grapple not with mere language models, but with the very architecture of synthetic cognition itself—a theatre where only those forged in the fires of foundational science dare tread. The future belongs not to the charming impresario or the fearful prophet, but to the quiet architect who understands that reality, digital or otherwise, bends only to the will of mathematical truth"
Do you *really* think yours and mine sound similar?? Honest question, (this could be valuable for another article)
I do. There's parts in the above that I think are more OTT, and parts in the article I think are more OTT.
I'm not sure on it's value though - I'm only one reader. The main thing for me was just that it seemed a complete 180 from your usual style.
For example, I (personally) can't recall ever reading sentences like the below from your blog:
- "Their work echoed the same intellectual ambition and richness of spirit I might have felt toward Einstein’s relativity or the quantum mechanics of Schrödinger, Bohr, and Dirac"
- "He just keeps fleeing forward, chased by his own shadow—the shadow of a long string of mistakes, the latest of which was a disaster of Babelian proportions that, while negligible in consequence, left him exposed"
- The entire 'Nerve' segment
I wouldn't put too much weight in just my opinion though.
I pasted your past 5 articles into Sonnet w/thinking and asked it to compare: https://snipboard.io/fzXd9w.jpg
Believing Musk to be a positive force for the world after we've seen him cheer on AFG while doing Nazi salutes and getting Trump elected certainly clarifies what sort of world you are hoping for
"reading comprehension, not for the emotionally reactive"
Do you think that Sam Altman is more dangerous than someone who did a Nazi salute with perfect form at the inauguration?
Maybe you haven’t seen video of the salute and you’re just basing your opinion of the salute on articles that called it awkward hand waving that is being misinterpreted? Because it definitely 100% was a Nazi salute, no doubt whatsoever.
I’m not trying to do some gotcha thing here. I probably comment on your blog more than any single other reader lol. You’ve got my loyalty.
The exact phrase was “I think Musk is closer to being a positive force in the world”
I hear that closER, not literally a positive force, just closer to being one.
But you should know that if you’re saying that Musk is better for the world than Altman, that you’re suggesting that Altman is more dangerous than Nazi-ism.
That’s a pretty bold position to take. I could see how it might be true… AI could create a permenant dictatorship, p(doom) all that. I’d want to hear more about it from you, though.
Hey Geoffe, thanks for the question. Nazism is the secular version of the devil - except perhaps complete destruction of humanity I can't think of anything worse. Yes, I saw the vid. Elon Musk has grown severely unhinged and the salute was real and intentional. However, unhingedness may not be as problematic as master manipulation (i.e., signaling is not the same as doing). That's my reasoning at least! (That said, if Musk turns out to be a real hands-on nazi and aims to influence the political sphere on those terms, then that's a completely different scenario than what I'm assessing here.)
FWIW I don’t think Musk is a hands-on Nazi either. I think it was signaling too, and that his rationale was cynical, not ideological.
I think he did the salute to excite and titillate some of his worst allies and strike some serious, paralyzing fear into some of Trump’s enemies.
The whole first 100 days in office seemed like a propaganda assault more than anything else. Just a bunch of stuff that probably won’t hold up in court, done to throw people off balance and terrify them.
Their strategy seemed to rely on so much sacred cow desecration that when someone more liminal (eg the algorithmic bridge) says something fairly innocuous (eg “closer to positive”), it instantly creates a whole fucking issue that only serves Trump and Musk in the end. They desperately want attention and fear sent their way because it’s the root of their power and authority.
Anyway, way outside the scope of this blog.
Thanks for your explanation.
I had a feeling you’d get in trouble for that sentence. Hope this is the worst of it.
I'm used to this kind of trouble tbh hahaha. Anyway, glad you asked (not being American - ie, not having a special predilection to any particular side - is both a blessing and a curse when it comes to these topics)
signing the secular version of the devil to his allies seems to presuppose some terrible (and correct) things about his allies, allies that he financially, politically, and via propaganda supports.
What does Sam do that rises to that level?
No you did not hide your preferences with the "closer to" qualifier.
i'm telling you now those are not my preferences but you can keep going
I'd love to see where Altman is the most dishonest player in SV. Musk's recent political moves for me totally disqualify him as a potential "good guy" - his latest behaviour is pure evil and pushing for Trump when you built your persona on saving the planet is as moronic as it is ironic.
I actually think that Altman has this true 360 vision - based on his time spent at YC and his other investments that none of the others have. And he does things with a lot of cool, like he's having fun with his colleagues - just look at the demos he did with the teams. While building a company from scratch with 500 million users.
Like you I have lots of respect for Hassabis, who is the only true genius of the lot.
Not sure he's the most dishonest player though I just... Don't trust him. The "hunch" is not a rhetorical figure but quite literally how my intuition speaks about him
I just listened to him this morning on https://www.youtube.com/watch?v=ctcMA6chfDY and I really enjoy the (apparent) lack of hubris.
In the first two mins of that interview you already see that he's effortlessly witty and that his conviction has always been strong!
I mean, it's all relative.
Deepmind has been pretty good on safety as well