OpenAI researchers warned board of AI breakthrough ahead of CEO ouster • Hacker News Board

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#501

Post by thaumaturgy » Thu, Nov 23, 2023, 4:57 AM UTC

huitzitziltzin wrote: post #358

Put this on my tombstone after the robots kill me or whatever, but I think all “AI safety” concerns are a wild overreaction totally out of proportion to the actual capabilities of these models. I just haven’t seen anything in the past year which makes me remotely fearful about the future of humanity, including both our continued existence and our continued employment.

The clear pattern for most of human history is conflict between a few people who have a lot of power and the many more people that are exploited by those few. It should be obvious by this point that the most probable near-term risk of AI development is that wealthy and influential groups get access to a resource that makes it cheap for them to dramatically expand their power and control over everyone else.

What will society look like when some software can immediately aggregate an enormous amount of data about consumers and use that to adjust their behavior? What might happen when AI starts writing legislation for anybody that can afford to pay for it? What might AI-generated textbooks look like in 50 years?

These are all tools that could be wielded in any of these ways to improve life for lots of people, or to ensure that their lives never improve. Which outcome you believe is more likely largely depends on which news you consume -- and AI is already being used to write that.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#502

Post by Davidzheng » Thu, Nov 23, 2023, 4:59 AM UTC

alchemist1e9 wrote: post #461

Davidzheng wrote: post #436

Earlier quoted context omitted.

I will bet ANY amount of money that 30% of current jobs will be displaced with AI advances in the next twenty years.

Darn well I really was hoping my children and grandchildren could continue my wonderful data entry career but OCR ruined that, and now they can’t even do such meaningful jobs like read emails and schedule appointments, or do taxes like an accountant. What meaning will they have in life with all those ever so profound careers ruined!! /s

Fair! But I'm just reminding the comment above that continued employment is not really guaranteed for most jobs which are mostly mental.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#503

Post by muskmusk » Thu, Nov 23, 2023, 4:59 AM UTC

cduzz wrote: post #316

I was talking to my (12 year old) son about parts of math he finds boring. He said that he thinks absolute value is absurdly easy and extremely boring. I asked him if there was anything that might make it more interesting, he said "maybe complex numbers". So I asked him "what would the absolute value of i+1 be?" he thinks for a little bit and says "square root of 2" and I ask him "what about the absolute value of 2i…

If you ask Ilya Sutskever he will say your kids head is full of neurons, so is LLMs.

LLMs comsume training data and can then be asked questions. How different is that to your son watching YouTube and then answering questions?

It's not 1:1 the same,yet, but it's in the neighborhood.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#504

Post by alienicecream » Thu, Nov 23, 2023, 5:00 AM UTC

Jugglerofworlds wrote: post #444

adriand wrote: post #318

Earlier quoted context omitted.

It strongly reminds me of the UFO stories that were all the rage a few months ago, and that military guy who testified before Congress about what he had heard from other people. Did any of that pan out? It seems not.

I follow the UFO community, that stuff is still going on. The military man, David Grusch was recently on the Joe Rogan podcast where he talked about steps moving forward. The next big thing coming down the pipeline is ensuring that "The UAP Disclosure Act of 2023" proposed by Senate Majority Leader Chuck Schumer passes the house. Yes this is the real name of the act, and has already passed through the Senate as part…

It's no fun if you don't get strung along for a few years.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#505

Post by abeppu » Thu, Nov 23, 2023, 5:01 AM UTC

wbhart wrote: post #491

I feel very comfortable saying, as a mathematician, that the ability to solve grade school maths problems would not be at all a predictor of ability to solve real mathematical problems at a research level. The reason LLMs fail at solving mathematical problems is because: 1) they are terrible at arithmetic, 2) they are terrible at algebra, but most importantly, 3) they are terrible at complex reasoning (more specifica…

This comment seems to presume that Q* is related to existing LLM work -- which isn't stated in the article. Others have guessed that the 'Q' in Q* is from Q-learning in RL. In particular backtracking, which you point out LLMs cannot do, would not be an issue in an appropriate RL setup.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#506

Post by markisus » Thu, Nov 23, 2023, 5:03 AM UTC

lovecg wrote: post #389

CaptainNegative wrote: post #282

Earlier quoted context omitted.

Here's a simpler example that works. Q: Prove that if cos(x) is an integer then so is sin(x). A: The statement "if cos(x) is an integer, then so is sin(x)" is not true in general. To understand why, let’s consider the properties of sine and cosine. [...] It had trouble when the problem domain is different from the usual one the same keywords typically fall into.

It does just fine for me (I’m impressed): You: Prove that if cos(x) is an integer then so is sin(x) ChatGPT: To prove this statement, let's first establish some basic facts about the sine and cosine functions: 1. *Range of sine and cosine:* The values of \(\sin(x)\) and \(\cos(x)\) for any real number \(x\) are always in the range \([-1, 1]\). 2. *Pythagorean Identity:* For any angle \(x\), \(\sin^2(x) + \cos^2(x) =…

Interesting. Which version did you use? My experiment failed (GPT 3.5)

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#507

Post by PaulDavisThe1st » Thu, Nov 23, 2023, 5:03 AM UTC

jjeaff wrote: post #458

PaulDavisThe1st wrote: post #268

Earlier quoted context omitted.

Except that there's a fairly large body of evidence that persuasion is of limited use in shifting political opinion. So the persuasion would need to be applied to something other than some sort of causative political-implication-laden argument.

When you say persuasion, are you referring to fact based, logical argument? Because there are lots of other types of persuasion and certainly some work very well. Lying and telling people what they want to hear without too many details while dog whistling in ways that confirm their prejudices seems to be working pretty well for some people.

    * **that confirm their prejudices** *

(my emphasis)

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#508

Post by nijave » Thu, Nov 23, 2023, 5:03 AM UTC

wbhart wrote: post #491

I feel very comfortable saying, as a mathematician, that the ability to solve grade school maths problems would not be at all a predictor of ability to solve real mathematical problems at a research level. The reason LLMs fail at solving mathematical problems is because: 1) they are terrible at arithmetic, 2) they are terrible at algebra, but most importantly, 3) they are terrible at complex reasoning (more specifica…

ChatGPT (3.5) seems to do some rudimentary backtracking when told it's wrong enough times. However, it does seem to do very poorly in the logic department. LLMs can't seem to pick out nuance and separate similar ideas that are technically/logically different.

They're good at putting things together commonly found together but not so good at separating concepts back out into more detailed sub pieces.

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#509

Post by muskmusk » Thu, Nov 23, 2023, 5:03 AM UTC

wbhart wrote: post #491

I feel very comfortable saying, as a mathematician, that the ability to solve grade school maths problems would not be at all a predictor of ability to solve real mathematical problems at a research level. The reason LLMs fail at solving mathematical problems is because: 1) they are terrible at arithmetic, 2) they are terrible at algebra, but most importantly, 3) they are terrible at complex reasoning (more specifica…

Friend, the creator of this new progress is a machine learning PhD with a decade of experience in pushing machine learning forward. He knows a lot of math too. Maybe there is a chance that he too can tell the difference between a meaningless advance and an important one?

Re: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster

#510

Post by somesortofthing » Thu, Nov 23, 2023, 5:03 AM UTC

dmichulke wrote: post #493

jhbadger wrote: post #380

Earlier quoted context omitted.

Exactly. The rational fear is that they will automate many lower middle class jobs and cause unemployment, not that Terminator was a documentary.

By this logic we should just forbid the wheel. Imagine how many untrained people could work in transport and there would always be demand. So why did the wheel not result in mass unemployment? And factories neither? Certainly it should have happened already but somehow it never did...

Past technological breakthroughs have required large, costly retools of society though. Increasingly, those retools have resulted in more and more people working in jobs whose societal value is dubious at best. Whether the next breakthrough(or the next five) finally requires a retool whose cost we can't afford is an open question.