OT: Are we on the cusp of an “A.I. bubble?”

[zmalwo]

A famous economist Frugman once said the internet will be no more than a fax machine by 2005. He won the Nobel prize for economics too.

[rounder]

I am just a casual user (am retired so who cares). We have a Sonos and a few other Alexa devices. I am impressed by the questions that Alexa can answer instantly (what was the score of the Orioles game, what was the answer to 25 down of today's crossword puzzle, what is a proteus anguinus salamander) and others.

Not sure that my job could have been replaced by artificial intelligence but I imagine that a lot of them will be.

[prototoast]

Originally Posted by zmalwo

A famous economist Frugman once said the internet will be no more than a fax machine by 2005. He won the Nobel prize for economy too.

Mr Krugman's nobel prize was for work predating, and largely unrelated to anything he ever writes about as an opinion writer for the New York times.

The reality is that doing academic writing is really hard. Have to find lots of evidence, develop rigorous models, and do so well enough to satisfy the editors and peer reviewers.

In contrast, writing opinion pieces for the New York times largely just requires making stuff up with minimal, if any, research required. It's not a surprise that someone would find the latter career easier and therefore preferable to the former, but he was pretty good at the former.

[RFC]

Originally Posted by prototoast

Mr Krugman's nobel prize was for work predating, and largely unrelated to anything he ever writes about as an opinion writer for the New York times.

The reality is that doing academic writing is really hard. Have to find lots of evidence, develop rigorous models, and do so well enough to satisfy the editors and peer reviewers.

In contrast, writing opinion pieces for the New York times largely just requires making stuff up with minimal, if any, research required. It's not a surprise that someone would find the latter career easier and therefore preferable to the former, but he was pretty good at the former.

Bravo. I like Krugman for both types of his writing. I am an academic writer and opinion writer, but not in his class. I view my opinion writing as experimental.

Likewise, I view my playing with AI as experimental.

[reuben]

Originally Posted by prototoast

In contrast, writing opinion pieces for the New York times largely just requires making stuff up with minimal, if any, research required.

Sounds like the perfect job for AI.

[martl]

All those party tricks like generating "art" pictures, or writing essays, is all fun but it doesn't really justify the effort. Problem is, these implementations like ChatGPT are quite good at generating answers that look and feel as if a human had written them, but since it is impossible to trace back where and how anything came from, the result can be anything from absolutely 100% correct to complete to utter nonsense. Everything coming out of an AI needs to be checked and double checked.
Where it is really good at is as acting as some kind of smart search algorithm -this is already used by browsers like Edge- and/or, for example in research, at looking at larger amounts of data and finding relations, correlations etc. Here it is beneficial that the human factor (bias) is not present -although i think OpenAI etc. start to mimick that, too...

AI is hideously expensive both to develop and to run. Anthropic founder Amodei estimates the cost to develop a "large language model" from 1 Bio $ upwards. To build and run an AI, you need people who understand how to do that, these are neither many of about, nor do they come cheap. There needs to be a lot of benefit in it to justify the investment. Cant get that kind of return from an automated chatbot on your customer relations website.

[redir]

Originally Posted by peanutgallery

in some ways ai could be considered an e-bike for the intellect

:d:d:d:d:d:d

[ridethecliche]

Blowing my mind that this NVDA rally is still going. I expected a bit of a slowdown, but as someone that's considered jumping into tech stonkz a few times as a hobby investor, it just keeps blowing my mind that things explode the second it feels like the bubble is about to pop.

This has been a wild ride.

110's fall of 2022 to over 1000 today.

[echelon_john]

They sell oil rigs. There will certainly be more competition than there is today for Nvda but as more and more private enterprises seek to develop their own internal ai platforms demand will remain strong into at least 2026.

Fun when it was just a graphics chip and the stock sold for $4/share.

Originally Posted by ridethecliche

Blowing my mind that this NVDA rally is still going. I expected a bit of a slowdown, but as someone that's considered jumping into tech stonkz a few times as a hobby investor, it just keeps blowing my mind that things explode the second it feels like the bubble is about to pop.

This has been a wild ride.

110's fall of 2022 to over 1000 today.

[julian3141]

Originally Posted by verticaldoug

I was commenting on the IVY League students. Incredibly underwhelming.

The reason alot of systems are called co-pilots, you can't trust and shouldn't trust the AI to pilot. It's a bit like sitting in the backseat and letting your tesla drive you around. You eventually will have to grab the wheel.

yeah, it was pretty bush league. I always had questions about the ethics of my cohort. So many of these kids are so competitive and type A I would not be surprised by any let's say bending of the rules.

[m_sasso]

An interesting take by Michel Barry on AI comming to riders cycling data and race performance: https://cyclingmagazine.ca/sections/...g-even-faster/

[echelon_john]

I think it even goes beyond what Barry is saying about data. The analysis of video in sport in real time can tell when an opponent is tired, if their movements might indicate an injury, all sorts of things that can be used to shape tactics. I wouldn’t be surprised in the least if it were employed in high stakes poker to analyze facial expressions and link them to cards/betting behaviors over time.

The deductive power of these tools is way more impressive than the generative power currently.

And on the topic of AI bias, which is deservedly getting a lot of attention, my strong opinion is that ai tools simply don’t need to take into account information that could drive bias because there’s simply so much other behavioral data available. Iow, the ai doesn’t need to see if you’re in a high-rent or low-rent zip code or look at other proxies for identity which can feed bias because there’s lots of other ways for it to identify propensity for actions, etc.

Originally Posted by m_sasso

An interesting take by Michel Barry on AI comming to riders cycling data and race performance: https://cyclingmagazine.ca/sections/...g-even-faster/

[Fat Cat]

I think that it would be interesting for someone to ask a generative AI what is the best way to kill an AI and see if it comes after him.

[jimoots]

Originally Posted by verticaldoug

I think you are using the wrong metric. When you say 'Bubble' you are really speaking about valuation in an economic sense. In 1999, the internet craze was a bubble, and the market did crash and a lot of bad ideas went out of business. However, the internet proved really useful and has continued to grow to valuations which are really quite staggering when compared to 1999. The Internet was really useful before the iphone (or smartphone) in general. But the ability to have access to the world wide web while mobile has really changed things.

So is AI a bubble. Yes, probably since the same lack of discipline for investing in any hairbrained startup seems to be upon us. But will AI prove extremely useful and is here to stay? Without a doubt. In 15 years, will we look back and be staggered on how far we have progressed, we are just beginning to see the build out of the software stack which will really improve the useability of GenAI. Just like in 1999, when the crash comes for AI, you need to take a deep breath and buy some of the survivors because those will continue to grow.

Yes this is exactly where I go. The bubble is in valuation, not the tech.

It's like picking pets.com or Amazon in 1999.

In 20 years time there will be a handful of winners, lots of losers, and on the whole we are likely to be more productive as consequence.

And on that, some winners may also be losers. Amazon with their Just Walk Out Technology is one.

[verticaldoug]

Originally Posted by jimoots

Yes this is exactly where I go. The bubble is in valuation, not the tech.

It's like picking pets.com or Amazon in 1999.

In 20 years time there will be a handful of winners, lots of losers, and on the whole we are likely to be more productive as consequence.

And on that, some winners may also be losers. Amazon with their Just Walk Out Technology is one.

Yes, but Just Walk Out was a hand wave, and not real. It deserved to fail and for Amazon, that was really a failure of process. I am shocked no one said 'this sounds like a bad idea'.

But have you tried Uniqlo's throw in the bin checkout? It's great and AI not needed.

https://www.forbes.com/sites/jonpico...ny-created-it/