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

Where does next-token prediction leave us?

pop.rdi.sh

Post reply on HN
Print view View original

211–220 of 221 posts

Re: Where does next-token prediction leave us?

#211

Post by aembleton »

amelius wrote: post #206
aembleton wrote: post #159

Earlier quoted context omitted.

> Now I feel something has been taken away and has no value. Did you feel similar when Wikipedia was created or Google or when you first got access to the internet? All of these tools have made digging into subjects and learning easier. I find the same with AI. I love it when a random thought pops into my head and I can explore it with an AI such as Gemini. Then ask it for the sources it used so that I can read furth…

> AI is just a tool You are forgetting that AI can also use tools. And those AIs can be used as tools as well by other AIs. The moment you start working on something, someone else is already automating the exact thing you are doing now.

> The moment you start working on something, someone else is already automating the exact thing you are doing now.

Good. If a machine can do my job then I can work on something more interesting. Perhaps a more interesting problem to fix is having the people working on something getting told about the automation.

Re: Where does next-token prediction leave us?

#212

Post by karlgkk »

aembleton wrote: post #211
amelius wrote: post #206

Earlier quoted context omitted.

> AI is just a tool You are forgetting that AI can also use tools. And those AIs can be used as tools as well by other AIs. The moment you start working on something, someone else is already automating the exact thing you are doing now.

> The moment you start working on something, someone else is already automating the exact thing you are doing now. Good. If a machine can do my job then I can work on something more interesting. Perhaps a more interesting problem to fix is having the people working on something getting told about the automation.

> If a machine can do my job then I can work on something more interesting

Oh my god, you don't get it.

The more interesting thing? Someone else is already automating doing that exact interesting thing you are doing now.

All the way down.

Re: Where does next-token prediction leave us?

#213

Post by keybored »

overgard wrote: post #182
dyauspitr wrote: post #57

Earlier quoted context omitted.

[flagged]

"want to see where it goes". Ok, what if it goes somewhere awful? A "paradigm shift" could easily just mean "corporate feudalism". It's not "small minded" to consider that not all change is good, and skepticism isn't a bad personality trait.

“Some of you may die but that’s a risk I’m willing to take.”

Then the peasant took that as an inspirational speech.

Re: Where does next-token prediction leave us?

#214

Post by futune »

Npovview wrote: post #27

People have bargaining power through their military services. If even that job is taken by AI, there is truly no recourse for the people left at all. These are the sentiments of Yuval Noah Harari.

The first person videos of drones hunting terrified soldiers coming out of Ukraine seem relevant. Apparently the last moments of the hunt are pretty much automated now, which increased the kill-rate considerably.

Re: Where does next-token prediction leave us?

#215

Post by kqr »

brabel wrote: post #100
kqr wrote: post #56

Earlier quoted context omitted.

As of today, "minor expert" is the wrong way to phrase it. The flagship LLMs today are at best at an initiate or apprentice level in every field.[1] This is not meant as a derisive remark – having something at hand that is initiate level in every field is remarkable and useful. But it's nowhere near expertise anywhere. [1]: https://entropicthoughts.com/stop-using-junior-and-senior

You seriously think current LLM is just at apprentice level in programming? It can write stuff one shot that I’d expected even some experts to struggle to do even with ample more time allowed.

Yes. The definition of the apprentice/journeyman boundary is that a journeyman can do a day's job unsupervised. LLMs cannot do that.

Re: Where does next-token prediction leave us?

#216

Post by brookst »

mlmonkey wrote: post #181
jack_pp wrote: post #39

Earlier quoted context omitted.

It costs between 4k and 10k if you break your leg and call an ambulance in the US. Forget medical advances, it costs you a car to get a cast

Not if you're homeless in, say, San Francisco.

Also not if you’re the president.

But generally we’re talking about the life experience of the 330m people who are not outliers in our society.

Re: Where does next-token prediction leave us?

#217

Post by brookst »

selcuka wrote: post #209
brookst wrote: post #174

Earlier quoted context omitted.

I don’t really find it fruitful to speculate about motives in complex systems. The claim was that average people don’t benefit from medical advances, which seems clearly untrue. Now, if we’re shifting to why advances are often generally available, sure, there are lots of plausible theories that may or may be true, in whole or part, and which we’ll never prove one way or the other. Believe what makes you happy, or ang…

> Now, if we’re shifting to why advances are often generally available But it's not me who is shifting, right? The claim was: > Science and technology, I feel, has always had a certain apathy towards the plight [of] the people at the bottom rungs. not > The claim was that average people don’t benefit from medical advances My point is to reject the claim that the high availability of vaccines is a counter-example. It'…

You’re still getting at motives rather than facts, which means it’s unfalsibisble.

If you’re going to say that fact X is not a counter/example because [subjective, speculative reasons] then I’m confident you’ll never find a fact that contradicts your beliefs. There will always be some “oh, well the bridge collapsed, but it wasn’t bad engineering it was political reasons, so it doesn’t count as a collapse.”

Re: Where does next-token prediction leave us?

#218

Post by brookst »

slopinthebag wrote: post #81
brookst wrote: post #23

Earlier quoted context omitted.

It’s not coding that’s being democratized, it’s the ability to create games, tools, etc, that require coding. Suppose it became possible to buy a Ferrari (not that new hideous one) for $5k. That would be democratizing Ferrari ownership: far more people could do it. I’m sure there would be investment bankers who would complain it devalues their hard work (I am not a fan of those people, but it is notoriously long hour…

If that happened, Farrari's would simply lose their value and something more rare would take it's place. The scarcity is the point. Same with art, if you "democratise" it somehow humans will seek out a new form of artistic expression that is not accessible to people with a $10 midjourney subscription. And that is OK, good actually.

Perhaps to some degree, but there is intrinsic value to a Ferrari if you enjoy driving. That doesn’t change with scarcity.

But I agree, and even celebrate, that a hypothetical future with abundance will be bad for speculators and the luxury-obsessed, and great for people who enjoy the first-order benefits of products.

Re: Where does next-token prediction leave us?

#219

Post by selcuka »

brookst wrote: post #217
selcuka wrote: post #209

Earlier quoted context omitted.

> Now, if we’re shifting to why advances are often generally available But it's not me who is shifting, right? The claim was: > Science and technology, I feel, has always had a certain apathy towards the plight [of] the people at the bottom rungs. not > The claim was that average people don’t benefit from medical advances My point is to reject the claim that the high availability of vaccines is a counter-example. It'…

You’re still getting at motives rather than facts, which means it’s unfalsibisble. If you’re going to say that fact X is not a counter/example because [subjective, speculative reasons] then I’m confident you’ll never find a fact that contradicts your beliefs. There will always be some “oh, well the bridge collapsed, but it wasn’t bad engineering it was political reasons, so it doesn’t count as a collapse.”

What I said is not subjective/speculative. It is not even unfalsifiable.

To restate, COVID vaccine can't be used as an example of "science/technology's sympathy towards the bottom rungs". They were only accessible because governments purchased those vaccines from science/technology companies and gave away for free. Companies developed it to make profit.

If the governments didn't purchase billions of doses at a subsidised price, most people wouldn't be able to afford it [1]. I don't see how this conclusion is subjective, speculative, or unfalsifiable?

[1] https://www.sbs.com.au/news/article/covid-19-vaccines-what-i...

Re: Where does next-token prediction leave us?

#220

Post by deterministic »

> Non-technical middle managers who have not written a line of code in their lives, now feel that the biggest obstacle between them and greatness has lifted.

Or it will turn out we no longer need middle managers because of AI.

Side note: I am pretty sure most companies could fire 50% of middle management and see a dramatic productivity improvement. With or without AI.

Post reply on HN