Hacker News Reader

Comment by endisneigh | Hacker News Reader

These sort of articles are so boring. Give me a date, or a year. Otherwise Doomers will continue to be vague and when something does happen, maybe, they’ll all come out of the woodwork saying I told you so.

Until then I’ll enjoy the unprecedented prosperity we’re enjoying here on earth.

It’s the same with stock market doomers, housing doomers, etc etc. no one ever gives a specific date. People just make unfalsifiable claims.

View on HN

I mean, your parent's doctor doesn't tell you a specific date your parent is going to die from cancer. Why would anyone else be able to predict dates that depend on more complex factors?

Doctors do give pretty tight estimates on expected life expectancy given the stage of the cancer, though.

That depends heavily on the type of cancer.

It's happening right this second. Right now the American West is suffering from incredible drought, right now there are water problems there. Right now heat waves in India are causing crop failures meaning then are considering to stop the export of wheat at the same time that a war for resources in Ukraine is cause them to reduce wheat exports.

I've been a "doomer" so to speak since 2016. I'm shocked at how many things I was concerned about have come to pass. If you were to describe the world of today to someone from 1996 they would be horrified and think you were full of nonsense.

But denial is a hell of a drug, so I'm guess you and many other will still not see it no matter what happens. There was a documentary on the sinking island of Tangiers. The mayor of the town (and many residents) remained rabid climate change deniers even as they witnessed their own home be consumed by the sea. Watching that made me realize that nothing will for people to realize what's happening when that realization comes at such a high mental cost.

> If you were to describe the world of today to someone from 1996 they would be horrified and think you were full of nonsense.

You think an average person in 1996 would be horrified of the world today?

I know the 1996 version of me would be horrified of the world today. I also know I'm not very representative, so take that as you will.

> Until then I’ll enjoy the unprecedented prosperity we’re enjoying here on earth.

Who is the "we" who gets to enjoy this? Us here on this site for the most part certainly but who else, who is not included?

How's that heat wave in india going btw. Unprecedented prosperity indeed.

Do you think India has historically been better off than it is now?

Do you believe more than 10% of countries (20 in about 200) are worse off in 2020 than 1920?

To me, this does seem like unprecedented prosperity — on a global scale.

Wasn't on your list, but personally: Singularity 50/50 by 2025. Past 2028 you get to say "I told you so."

Three years? That’s a stunningly optimistic (pessimistic?) timeline…

What can I say, PaLM is scary.

The hand-picked examples are pretty scary, but I'm also reasonably sure that AGI is not going to come (directly) from LLMs no matter how big they get.

Reasoning: https://medium.com/@emilymenonbender/on-nyt-magazine-on-ai-r...

Seems too focused on the NYT article? I don't think this because the NYT said so, I think this because PaLM's chains-of-thought approach seems to match how my own brain works. I think LMs have huge intelligence overhang atm that is hidden by the way they are forced to respond. If we figure out how to train LMs with propagation through hidden chains of thought, we'll see a huge jump in capability, and also this is the obvious next step that the industry will aim for.

I’m betting 2025-2030 as well.

We have quantum AI math developed and photonic accelerators are currently in development. The moment those work reliably, machine learning will hit an exponential curve — particularly since those photonic circuits are already partially designed by AI.

The feedback loop of quantum-accelerated AI refining its own hardware is one breakthrough away.

Stop reading so much popsci, it is bad for thought.

This is an interesting criticism. I grew up on wired magazine and sci-fi. Even then I recognized the dangers of shallow knowledge and maybe I can’t avoid it even now. But I’m actively working with GPT3 and I must say it is startling. I’d say we’re a few years from automating most knowledge tasks (I created a tool that I chose not to release because it was too powerful). We could be a decade from general purpose artificial intelligence (or it might not be possible at all. For a fun thought experiment ask GPT3 if it is truly conscious). The answer it gave me was surprisingly coherent wise and useful. From there (AGI) we could be minutes away from artificial super intelligence. That starts to look a bit like a point pst which standard prediction breaks down. But like you say, pop sci might be hindering rational thinking on this score.

I used to work in physics, which seems to attract a lot of people with shallow knowledge, and ML seems to be going in a similar direction.

To be clear, I'm actually quite bullish on AI (well, bullish on it ultimately happening, not necessarily on how it will impact humanity). I was mostly reacting to the timeline for photonic quantum computing rather than on ML advancement, actually.

Language modeling is extremely impressive. I've been saying since 2017 that modeling language is basically modeling the world/intelligence, but at the time most people did not seem to believe that. People still don't really seem to believe it, but I think the smarter among us are beginning to now that they see GPT-X & PaLM. I think there is still a substantial gap from "good language model" to "embodied agent taking actions in the world" that we have not bridged and will be potentially pretty hard to, more than mere "minutes."

> will be potentially pretty hard to, more than mere "minutes."

This makes it sound like you’re arguing against something no one said — which makes it a weird thing to quote.

A quote from the GP

> The answer it gave me was surprisingly coherent wise and useful. From there (AGI) we could be minutes away from artificial super intelligence

What am I missing? I assumed of course that they didn't mean literal minutes but more metaphorically, but my reply was also using it in the metaphoric sense - it could be a potentially heavy lift to go from point A to point B.

e: I guess I am missing that they are conflating GPT with AGI and saying it will be minutes from AGI to artificial super intelligence in the literal sense. That, I actually do agree with, but I don't think GPT-X or PaLM or LMs in general qualify as such.

I read it in a different way, where we’d go from “plays chess like a bad human” to “plays chess superhumanly”, applied to general knowledge synthesis tasks — something that would be a rapid change once it was able to synthesize general knowledge answers.

Upon reflection, my comment was too critical for having a different read than you did.

I’m sorry that you’re ignorant.

Microsoft QuArC has published a number of papers on QAI, and I both work on photonics and discuss with other companies who do the accelerators being developed. Accelerating AI with photonics is an active area of development.

I think you’re projecting that your knowledge of this is only based on pop-sci.

There are a ton of unsolved problems in both ML and quantum computing with photonics.

We are so far away from being able to do this with photonic quantum computers that I am really skeptical that you actively work in this field.

FWIW, I know quite a lot of people involved in AMO physics (which involves related techniques) and work in ML myself.

2025-2030 is way too soon for these things to come to fruition.

I guess you know less talented people then.

We’re not more than a single breakthrough from being able to make photonic accelerators — and there’s already worked out algorithms to accelerate ML once those circuits are available.

I’m confident in 2025-2030.

This is one of the ~top two universities in the country for AMO research, so I don't think so.

Such false bravado, not going to keep replying.

> Such false bravado, not going to keep replying.

Yawn — you’re the one who came in with ad hominems and trying to tell other people what they work on, lol

I’m sorry your friends are ignorant and you don’t know what I do.

Reply on news.ycombinator.com