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

My AI Timelines Have Sped Up (Again)

alexirpan.com

Post reply on HN
Print view View original

31–40 of 99 posts

Re: My AI Timelines Have Sped Up (Again)

#31

Post by lovecg »

vallode wrote: post #7

I'm no AI expert, I don't have a lot of domain knowledge but what irks me about reading through this entire blog post is the idea that we are hinging on a definition of AGI that apparently accounts for 95% (!) of "economically valuable work" while basically not talking _at all_ about the trades. No mention of construction, manual labour, mechanics, welders, painters, nothing! Is the author trying to say that all that…

Yeah by that definition we already eliminated 95% of economically valuable work from 1700s. Very few people do farm work nowadays.

Re: My AI Timelines Have Sped Up (Again)

#32

Post by gmerc »

siliconc0w wrote: post #20

I'm an "overhang" bear, while incremental progress will be made - it'll come at too high of a cost. The larger, slower, more expensive models have limited applications(mostly fine-tuning the economical ones) so there will be ultimately less investment. Unless there is a new unlock, we'll be left with mediocre 3.5 models for most real-time apps and excitement will fade.

ChatGPT is one giant training data acquisition project. Including pro. it’s painfully obvious when you see that any useful usecase eventually moves to the API (GPT 4 web/app flat out refuses to code more than a function or two at a time now; it had no issues with multiple files before).

Everyone is providing the greatest training set you can now find (since all other venues for training data will gradually disappear behind walls and fees)

Re: My AI Timelines Have Sped Up (Again)

#33

Post by oblio »

anonzzzies wrote: post #11
vallode wrote: post #7

I'm no AI expert, I don't have a lot of domain knowledge but what irks me about reading through this entire blog post is the idea that we are hinging on a definition of AGI that apparently accounts for 95% (!) of "economically valuable work" while basically not talking _at all_ about the trades. No mention of construction, manual labour, mechanics, welders, painters, nothing! Is the author trying to say that all that…

I think the idea of most is that at the time we have AGI, we will robots designing robots that will ‘fix’ AI for the trades.

Ah, yes, that second most famous AI operator after the tensor, the handwaver.

Re: My AI Timelines Have Sped Up (Again)

#34

Post by spywaregorilla »

vallode wrote: post #7

I'm no AI expert, I don't have a lot of domain knowledge but what irks me about reading through this entire blog post is the idea that we are hinging on a definition of AGI that apparently accounts for 95% (!) of "economically valuable work" while basically not talking _at all_ about the trades. No mention of construction, manual labour, mechanics, welders, painters, nothing! Is the author trying to say that all that…

The more technology can do, the less the workers do, the less workers get paid. Being a servant to people with money is better than being part of their furniture though.

Re: My AI Timelines Have Sped Up (Again)

#35

Post by croes »

Is really a sped up?

The target of 95% is still 2070, the 50% is still 2045, means the progress between 2035 and 2045 is now less than before.

So the timeline has sped up and slowed down at the same time

Re: My AI Timelines Have Sped Up (Again)

#36

Post by gumballindie »

kiba wrote: post #21
ChatGTP wrote: post #12

Earlier quoted context omitted.

The belief is that once the AGI is here, the possibilities are endless for disrupting absolutely every aspect of our life and society, the end of human history is what I've hear Open AI folks say. There will be zero work, humans will be completely useless economically and we'll just have to hope that some how, all of this wealth is fairly distributed and we don't lose our minds from feeling completely useless. What y…

We don't know what will be at the end of this. It seems that the OpenAI folks are endlessly pursuing the technology of AGI without actually working on making sure that AGI benefit "all of humanity." They haven't even articulate what it means to benefit all of humanity.

OpenAI is a cult. I dont quite understand why they are given so much airtime other than wanting to depress tech wages and drive traffic. Their products dont work as advertised and are built on theft.

Re: My AI Timelines Have Sped Up (Again)

#37

Post by vallode »

tomNth wrote: post #25
vallode wrote: post #7

I'm no AI expert, I don't have a lot of domain knowledge but what irks me about reading through this entire blog post is the idea that we are hinging on a definition of AGI that apparently accounts for 95% (!) of "economically valuable work" while basically not talking _at all_ about the trades. No mention of construction, manual labour, mechanics, welders, painters, nothing! Is the author trying to say that all that…

"reading through this entire blog post" ? Is "The Physically Embodied Bear" "95%+ will include taking physical, real-world actions." not a mention of "construction, manual labour, mechanics, welders, painters" and "blue-collar workers" ?

The paragraph was my exact problem, hundreds of words but the entire world of AI interacting with the world physically is boiled down into an assumption that "by the time we get there, we'll get there". It just doesn't seem realistic? Especially with predictions/hunches like the ones found in the article.

Others seem to mention it is a sort of feedback look, once we get to 50% AGI I guess that AI itself can help us overcome the physical limitations. I'm sceptical that it is as simple as that, but at the same time can see it as very much possible.

How much do current AIs contribute to their own development? Likely quite a bit, but then again I don't know any AI researchers to ask.

Re: My AI Timelines Have Sped Up (Again)

#38

Post by woopwoop »

Imagine you learn about some new conjecture in mathematics, and you try to predict when it will be resolved. Your prediction says that there is a 50% chance that the conjecture is resolved in the next 25 years, and a 90% chance it is resolved in the next 50 years. So, conditional on the conjecture not being resolved in the next 25 years, there is an 80% chance the conjecture is resolved in the subsequent 25 years. Wouldn't that be a little strange?

Re: My AI Timelines Have Sped Up (Again)

#39

Post by realce »

kiba wrote: post #21
ChatGTP wrote: post #12

Earlier quoted context omitted.

The belief is that once the AGI is here, the possibilities are endless for disrupting absolutely every aspect of our life and society, the end of human history is what I've hear Open AI folks say. There will be zero work, humans will be completely useless economically and we'll just have to hope that some how, all of this wealth is fairly distributed and we don't lose our minds from feeling completely useless. What y…

We don't know what will be at the end of this. It seems that the OpenAI folks are endlessly pursuing the technology of AGI without actually working on making sure that AGI benefit "all of humanity." They haven't even articulate what it means to benefit all of humanity.

They think they'll live in a super-utopia and the rest of us will simply die away. I hope we can start to face the REALITY of this change now instead of these nonsense salves like "more types of jobs will be created" or "UBI will work" because that's not how biological beings given unlimited power actually behave. They crush their biological competition as swiftly as possible. That's what will happen. The folks at OpenAI think they'll be on the winning side and know if they're not, they will be destroyed by a competitor's God Machine and die like the rest of us. Their zealotry is just fancy self-preservation at the cost of all the other humans.
Post reply on HN