Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Sam Altman may control our future – can he be trusted?

newyorker.com

Post reply on HN
Print view View original

721–730 of 1001 posts

Re: Sam Altman may control our future – can he be trusted?

#721

Post by surajrmal »

LAC-Tech wrote: post #561
pesus wrote: post #461

Earlier quoted context omitted.

Thank you for sharing this. I 100% believe it, and it lines up with my experience with other people who came from similar backgrounds as Sam Altman - i.e. white, rich, privileged, and attending elite universities. I will disagree with one part - I do believe it is racism. Most will never admit it publicly, but if they think you're one of them, it often comes out rather quickly, especially when alcohol is involved.

[flagged]

Are Irish red and Italians olive? Most folks of color consider Jewish folk of a certain complexion white. Heck, some Mexicans are seen as white.

Re: Sam Altman may control our future – can he be trusted?

#722

Post by arduanika »

tootie wrote: post #431
ronanfarrow wrote: post #307

Earlier quoted context omitted.

I’ve really appreciated how substantive and polite the discourse here is, overall!

Not a question but just wanted to make sure you saw this: https://theonion.com/anyone-else-have-those-weird-dreams-whe...

Also this exclusive interview with the man himself:

https://theonion.com/the-onions-exclusive-interview-with-sam...

Includes gems such as:

Q: What informs your personal sense of morality?

A: Previous things I’ve gotten away with.

Q: Why did you decide to devote your life to AI?

A: I just saw so much suffering in the world that needed to be automated.

Re: Sam Altman may control our future – can he be trusted?

#725

Post by jcgrillo »

davebren wrote: post #683
arcfour wrote: post #500

Earlier quoted context omitted.

> I'm not even sure we're any closer to AGI than we were before LLMs. I mean this is very obviously untrue. It'd be like saying we aren't any closer to space flight after watching a demonstration of the Wright Flyer. Before 2022-2023 AI could barely write coherent paragraphs; now it can one-shot an entire letter or program or blog post (even if it's full of LLM tropes). Just because something is overhyped doesn't mea…

Point is that LLMs could be a local minima we are now economically stuck in until the hype wears off.

Or we could be stuck here for decades pending a breakthrough nobody alive today can even conceive of, or we could be compute limited by a half dozen orders of magnitude. Or it could happen next week. That's the nature of breakthroughs--you just can't have any idea when or how (or if) they'll happen.

Re: Sam Altman may control our future – can he be trusted?

#728

Post by ethbr1 »

embedding-shape wrote: post #656
globalnode wrote: post #555

Earlier quoted context omitted.

at this point even googles ai search results are better than gpt - obv. this is not for full programs but if you know what youre doing and just want a snippet, thats all you need.

Wild how different experience people can have. Both Google's models and Anthrophic's hallucinate a lot for me, even when I try the expensive plans and with web searches, for some reason, and none of them come close to the accuracy and hallucination-free responses of ChatGPT Pro, which to me still is SOTA and has been since it was made available. But people keep having opposite experiences apparently, I just can't mak…

Kagi (assistant.kagi.com) with Kimi K2.5 (their current default) has worked great for me in scenarios where the search result data is more important than the model.

I.e. what I used to use Google for and when I don't want an AI to overly summarize / editorialize result data.

Re: Sam Altman may control our future – can he be trusted?

#729

Post by pch00 »

mplanchard wrote: post #660
pch00 wrote: post #652

Earlier quoted context omitted.

> The New Yorker prefers insure to ensure. They have a unique house style. That's not a stylistic choice, it's just incorrect use of English.

Well that’s just, like, your opinion, man. https://www.merriam-webster.com/dictionary/insure

That M-W entry literally says they're different words with different meanings:

> They are in fact different words, but with sufficient overlap in meaning and form as to create uncertainty as to which should be used when.

> We define ensure as “to make sure, certain, or safe” and one sense of insure, “to make certain especially by taking necessary measures and precautions,” is quite similar. But insure has the additional meaning “to provide or obtain insurance on or for,” which is not shared by ensure.

Re: Sam Altman may control our future – can he be trusted?

#730

Post by kogasa240p »

pharos92 wrote: post #153

We focus these critiques far too much on the face rather than the underlying mechanics. Just like in politics, we critique the personality/politician yet the underlying system architecture evades it. Sam Altman clearly has a long history of nefarious activity. But the underlying threat posted by AI to society, the economy and human freedom persists with or without his presence.

>But the underlying threat posted by AI to society, the economy and human freedom persists with or without his presence.

Without Sam Altman the compute and improvements for LLMs to be a threat wouldn't have readily existed at all. He was the one who got the ball rolling because of his desperation (SVB collapsed right before the hype bubble started), ego, and quasi-religious desires.

Post reply on HN