Hacker News Reader

> you can get more done in an hour than a normal person can get done in a day.

Motion, not progress. So much talk of 100x and 1000x engineers, so little to show for it.

Faster autocomplete doesn't make an AGI. We don't redefine the airplane just because high-speed rail is fast enough for 90% of commuters.

your brain just used fancy autocomplete based on past patterns to write this post

prove me wrong(you can’t)

Autocomplete in the sense of grammar, but not in the sense of content.

how do you know?

Grammar takes basic building blocks (words) and combines them into novel constructions, so grammar is "autocompleted", not content (for example finnegans wake or jabberwocky)

So why do humans keep trying to make a better LLM when the LLM is generally intelligent and should be able to make its own smarter LLM?

But that's according to your definition of AGI. According to OP's definition, stochastic word generation is the pinnacle of intelligence, the place where we should "stop waiting for 'our' AGI". Too bad he stopped there for his own.

I understand why people, even technical ones, give in to the hype. I sometimes do too when I look at the agent spinning, talking to itself as if it's an internal monologue and coming up with actions with real side effects on its own.

But when they start blogging or writing about AGI, especially _if_ they are technical, I wonder why don't they go just one, or at most two layers deep and try to understand why this generally intelligent contraption sometimes happily executes `rm -rf /` and then nonchalantly apologizes after the fact. They liken these slips to the same kind of mistakes humans do, but they are fundamentally different.

Human failure modes are predictable from their skill level. Humans also carry emotional baggage and they weigh the cost of the mistakes they bear depending on their privileges.

For an LLM "rm -rf /" is rare and unpredictable, but "I am sorry" after it is almost certain because they were trained on dialogs and they get better score for "I am sorry" than "that wasn't my fault".

There's no good definition. To some people it literally means "artificial general intelligence". If you agree that toddlers and dogs have general intelligence at a low level (and most do), that low bar has been met IMO.

Some people think it means "better than human intelligence in every way". We'll be able to find niches that humans are better at for a very long time, so that definition hasn't been met, but is getting increasingly pedantic.

Some people think it means "smart enough to improve itself independently, thus triggering the singularity". The frontier labs are using AI to help improve itself, but it's not working independently, yet, AFAICT. So that definition isn't met yet.

Psychosis is here

You forgot “Tell HN:”

"Peter, the cult is here."

A baby can identify most apples in the real world, after showing him just several pictures of them. And that's just the tip of general intelligence.

90%+ of humanity is dumb as rocks and wouldn't know "AI" or "AGI" from baby's first line of JavaScript.

Finally. Can we go eat something now?

I disagree with the first paragraph, agree with the second, and am not very hopeful about the third.

Lowest effort troll post I think I’ve ever seen boosted to the front page on HN.

I’m sure we top it soon though.

I wonder how many here or elsewhere care about Hinton (or others) predictions, not many I hope.

Happy to hear but I got to go because my forza download has just completed

AI is just a really fancy calculator that can run cron jobs

If AGI is here, the consequences of achieving AGI are largely disappointing.

1. You can't move goalposts to fit your narrative. AGI is here iff your definition of general intelligence is very limited. LLMs are nowhere near general human intelligence, so by many definition AGI is not here.

2. If AGI is here, so what? What are you saying we should do? What exactly are you selling? Why did you feel the urge to give us a heads up?

LOL.

On a more serious note: I don't think so.

I'm happily using the clanker all day every day, but that dumb piece of shit is definitely not an AGI.

> [...] like every other technology bad version 1.0 will give way to much improved version 2.0 [...]

Are you sure about that? We're talking about for-profit software here. Software tends to get worse and more user-adverserial as businesses think of new and insidious ways to ways to extract money from customers. There's no indication that any of the AI players have any greater aspiration that enshittifying the entire universe.

I just had meta's muse agent run my entire groups marketing campaign for my new app, it brought in 500 users in 2 weeks...it read the rules, advised on best posting practices on a per group basis and saved the notes and data for each group into a spreadsheet for future posts, then it created instagram videos and reels that are amazing and posted those to instagram and tiktok for me and it is creating a Kiva crowdfunding campaign for me....I don't know what else is required for AGI but that pretty much replaces the need for a personal assistant imho...

Tell HN: AGI Is Here