I’ve been seeing a lot of comments on social media describing AI as a bubble and hoping that the bubble pops soon.
This is the wrong reaction.
It is true that many companies (plus the CCP) are pumping huge amounts of money into AI and that a large number of those investments won’t pan out. It’s true that many companies are making shortsighted business decisions now they will later regret. That’s inevitable.
But AI scams are a thing now. Voice cloning and fake videos and vibe coding and slop accounts and spam websites and beauty filters and AI “agents” and so on… This is the “new normal.”
I can’t wait for the AI bubble to pop.
I also long for the days when Google Search was useful, propaganda more obvious, ignorance less articulate, when you could trust that the stranger you were arguing with online was at least nominally human!
But that world is gone. And nostalgia is not hope.
If you want a better future than this, you can’t simply wait. You have to make it. Demand AI regulation.
I think I read recently that it is more accurate to talk about an “AI investment bubble.” Just like how there was a bubble that popped with Web 2.0, but the technology that made Web 2.0 possible is still very much with us.
The reason to avoid AI from current models is … they just plain lie. This is also when it seems super technical and flash too! Test coverage that passes and writes the test “return true” or functions never implemented but called and “passing” etc things like that. Yes!
I think you’re correct bhante about it being a “wrong” reaction, in a relative and civic minded sense. As for what is “right” that’s debatable, one educated approach to evil is to infiltrate and organise, walk out with the docs and the goods “System Admins of the World, Unite” like assanges 2013 talk that proposed the chaos computer club (the largest tech association in europe) convert to a guild or union. The perspective then, and I think this is actually the long perspective of the (non-liberal) left, is to organise as workers knowing the system is alienating, it’s not like there’s consumer “choice”, working at Palantir? Good, keep working at palantir, and organise - so workers and civic control of the means of production. While there’s a belief that there’s some “choice” we have on where we work (that is somewhat true but also not - as society is so financialised that debt servitude is the normal experience ) there’s also a denial of choice we have to stop the evil we have control to stop as in - to shut down the factory or walk off the job, not load the ship, etc.
Another might be there’s also like the chinese hermit traditions where more than one person who was a civic minded institutional person saw the complete chaos and war and rampant evil and rather than being the stay and fight, ran to the mountains, a poet, contemplated, 25 years came back, suddenly a law-maker, etc pretty sure that’s a lesser known but worthy one our culture needs to adapt more hehe
Anyway I maintain AI is a bubble because its a subset of the world, and the world is a bubble lol xD
I can’t fathom how distant from the actual state of affairs in the world people who speak about “AI bubble” and especially that it is “gonna pop” appear to be…
YouTube has become flooded with deepfakes and, judging by the comments, massive numbers of people are not aware that they are being fooled and manipulated. The US political deepfakes are deeply, deeply concerning. Truth is becoming lost in the cacophony. It’s not a bubble.
Ok fathom this, diminishing marginal returns. I was shocked wasn’t just me and heard it in Palantirs last earnings call that they are away from that and think future is in the “edges” - what they mean is the client side prompting and stuff. Or fathom the exit of Yann Le Cunn from Facebook because he thinks theres no future in the large language model approach. Also consider all the adept programmers reporting mental health issues - like look at a video of Geohot - that’s not the LLMs becomign more powerful or capable - that’s the amount of technical debt and bull they have to debunk with while also restricting ability to be creative -
Yeah, that bit is not technology though thats the relationship between capital and the state.
But in terms of technology reminds me of all the other hypes “the information super highway” and its more like “wow this spreadsheet sure is efficient - formulas! automatic! - compared to those books” etc Also I think even people who designed their own models and released own self driving car hardware and widely considered “adept” programmer still don’t believe they may be doing anything to make things more productive (geohot) etc.
Also the effort to read some of the code they generate is huge comapred to human written stuff - Like imagine getting it to write an essay - then just feeding that essay back in again and again and again - theres some pretty crazy monstrosities out there.
Now consider this, that programs are often the business logic in institutions -
This is the problem - the “bubble” in the financial sense is when the markets and financial instutions and governments realise this bit - if they haven’t already factored that in - like in a sense - the model might be like Halliburton - they make the most when Cheney invade Iraq - in this case it would be “when all the institutions realise they now look like vibe code” then they can sell the fix, i guess
AI definitely has some useful purposes. Its incredible for language learning. My wife is a Chinese national, over there they embrace the good aspects of AI.
Firstly, it really depends on the language. I don’t know anything about the Chinese language and its structure so I can’t comment on Chinese specifically.
So much of the language learning material that has emerged over the past few years has been really bad, when it comes to European languages at least (specifically French). I was very enthusiastic about the possibilities of language apps, but the more I study the theory of language acquisition, the less I think so.
Apologies Bhante KR for taking this thread on a tangent. Language learning/translation is within the skope of D&D, so I hope that you’ll indulge me.
Well, if you ever actually learned French, you’d stop paying Duolingo, right? Their goal is subscription revenue, not French speakers, so they use game mechanics to get you hooked and feeling like you’re making progress while ensuring (pedagogically) you never actually get fluent.
Well, occasionally, Gemini ignores that my prompt is in English and replies to me in Thai, forcing me to get some practice reading Thai
I wouldn’t say “incredible” but there are ways of using LLMs that I’ve found modestly helpful.
Thai text doesn’t use punctuation or put spaces between words, which can be a bit overwhelming. So I sometimes ask the LLM to add spaces and English-style punctuation to a Thai text to make it more approachable for me.
And LLMs are better than Google Translate at translation into Thai. I still need to double check with a native speaker but it can give a good starting point for that conversation. “Does this sound natural: ____?”
Definitely in scope for the forum but maybe not in scope for this thread? Anyway, I’ll leave the moderating to the moderators.
That might make sense as someone outside of the United States, but one has to realize that the whole political cycle we are in at the moment was caused by the government bailout that happened in 2009 when the real estate bubble ended … the result was a country basically losing its mind and veering off the road into political chaos. At this point, our Congress is incapable of even passing annual budgets … a person would be nuts to invest with the assumption they will bail out a bunch of giant tech companies. I’m sure some politicians would love to do that, but there won’t be the political consensus to do that, much less agree on this year’s highway subsidies, etc.
Yeah, this is the real root of the problem isn’t it.
Absolutely everyone, including the AI companies, is asking for regulatory guidance and structure and guardrails to AI development. But US Congress is almost entirely deadlocked at this point as Gerrymandering and dark money and foreign influence campaigns and so on have turned Washington into a clown rodeo.
It’s not about ‘bailing out’ some companies, but rather about the true AI being an Ultimate Weapon (among many other things), and, therefore, an ultimate national security concern.
Therefore, it’s no longer a question of wild market capitalism, stock prices, fears of a market/finance crash/crisis, and general greed.
So all measures will be taken to maintain and accelerate AI progress and development, regardless of markets state.
And this is so worldwide, in Europe, Russia, China and America, so they don’t actually have a choice – only straightforward.
