Similar to what happened to NFT’s back in 2022
When the investment money stops flowing, that will be when you see AI disappear.
My company is in the middle of rolling out Microsoft Copilot or what ever it’s called. Here is the thing, CoPilot comes with Office and the company just spent a ton of cash on next year’s Office licenses. MS sold the people writing the checks that Copilot comes with Office and a project was created to specifically roll it out.
Fast forward another year, there will be polls of the employees how useful Copilot was in the past year… The employees will give their feedback. Then next year’s negotiations of Office licenses will include a discount for “removing” Copilot. As it’s not needed or even useful.
Then AI will disappear from our daily lives.
I’m guessing in a year, but honestly I think it might take up to 5 years as the money flowing into AI right now is insane and enough to keep it propped up for quite some time. I just hope it’s no more than another year.
I’ve read upon a particular prediction that I thought would be quite likely to happen. The bubble itself will likely burst, similar to the dot com bubble… but AI as a technology is likely here to stay, similar to the dot com bubble. The hope is that most of the current AI expertise/research will shift more toward academic institutions, who will primarily work on building scientific AI for research purposes. Similar to how foundation models/LLMs are already used in studying genetics/genomics and protein structures… And a few commercial AIs might eventually find a way to be profitable & stay on the market
Will suck a bit for the ML engineers who will have to find new jobs in academia and suffer massive pay cuts though
Are you talking about the neural networks that underly so-calldd “AI”?
NFTs are not a good comparison, because NFTs were only ever a gambling instrument, lacking any practical application.
Instead, the dotcom bubble is a better comparison. E-commerce didn’t go away after the bubble burst. Likewise, AI will continue to have applications and be part of the economy - as it also was prior to the LLM-driven boom. It’s just that some of the more bullshitty aspects will disappear, and the remainder will have more sensible market valuations.
NFTs actually do have a few practical uses, e.g. concert tickets, but the number is very small and the relevant industries are unlikely to adapt them anyways because capitalism
I agree with you, but concert tickets was a bad example, as that is easily centralizable. My go-to example to make people have a grasp of the type of problems this solves is games.
Imagine there was a store where you could buy games and activate them on any store you want to by only paying a small fee, for example you buy the game on that store and can activate the game on Steam, Playstation and Xbox only by paying the fees on each of them. This would allow people to essentially cross-buy games which is good for customers as it means they can move platforms or stores easier, it’s good for game devs because their games would get more views across the board and their percentage is clean, and it’s good for the stores because they get a clean cut of the cake when someone wants to activate a key.
Why can’t this be done with traditional methods? Because there’s no central authority on who owns a game, Valve has no reason to trust Sony and vice-versa. And even if you propose creating such centralized authority others might not trust it. A smart contract is transparent and auditable, requires no trust from any party and acts as a decentralized source of truth.
Using a decentralised system for concert tickets seems like an expensive way to take a centralised operation (the organiser sells the tickets and can manage resell / third parties) to force a distributed system (that will be less efficient since there is overhead ensuring distributed systems aren’t controllable by a single entity)
The benefit is less about decentralization in this case (although it would be nice because fuck corporations, especially Ticketmaster) and more about the use of cryptography. The blockchain and non-fungible aspect allows people to buy tickets secondhand without a doubt in their mind that the tickets are fakes or copies, all without the need for a corpo-approved marketplace which make the rich leeches richer.
You can use cryptography in a centralised system too. The issues with the fakes is more that its not supported / people try to work around the system. Why would a block chain backed system stay that way after they are entrenched? It’s layers on top and all those layers are separate layers that are trying to make money… that’s where there funding comes from and why they exist. Why would they not go down the same path since owning their layer gives that control. I have friends that were working exactly in that space and I saw nothing that made me think they would be any better besides currently being smaller
It’s going to be whatever new technology built on the shoulders of blockchain tech that works for this. Whatever they solve the efficiency problems with. Then it might have a fighting chance at being a currency, being used as provenance, etc.
I think AI’s got a similar issue with the agentic side of it. Still needs another big advance.
But… why? The centralised database uses less power so it’s a more efficient foundation. There are some situations where distribution might be worth that cost. A lot of those situations that people were excited about early on (a global currency / gold alternative) haven’t proven to be robust to the overhead (although speculative investments and transactions you don’t want tracked are working for some people)
All that said, how does a concert ticket ever fit into that foundation as a win… without bringing up a liberation type ideology?
No. AI can be useful. Right now they’re just throwing AI at the wall and seeing what sticks. Most of it will fail, but some of it will stick around.
There’s a boom?
Doesn’t “boom” usually also entail some sort of major profit being made? As far as I’ve heard AI is only costing everyone money, with no profit made anywhere.
Nvidia is making decent profit
https://www.macrotrends.net/stocks/charts/NVDA/nvidia/ebitda
Do they have an AI? I thought they were just selling hardware. That wouldn’t be an AI boom, it’d be more of a computer hardware boom.
They are the ones selling shovels
If you look at the bubble chart you’ll see that all the profit is just money moving around in a circle without any real income.
Yes it will. None of the current AI companies are profitable, have never been profitable and most likely will never be profitable. Their business model simply isn’t economically sustainable. What is being invested in AI is basically imaginary money in a circle-jerk of companies like Nvidia, Anthropic, OpenAI and others. This artificially bloats their value and there is zero chance in hell that the AI bubble will not burst like nothing ever has before.
This is not to say that LLMs and other fancy jargon generators cannot be useful. They absolutely can. What will probably happen is that people will realize that they are just fancy guessing machines without real understanding behind and learn to use them in that way. I hope at least.
Isn’t it just the training that costs a lot in comparison to running it? So if they stopped training new models and just sold their current ones it might be able to run at a profit from year to year, after a huge initial investment that may or may not ever be recovered.
Not really. The infrastructure to run language models is hardware wise demanding, consumes vast amounts of water (cooling) and electricity and requires frequent renewing and added capacity.
not in that way. its more like the dot com so its not going anywhere but its not going to take over everything. Its been a long time but at the time they talked about an end to brick and mortar. Which has kinda been happening, I mean kmart and sears are gone. Many brick and mortar do not have as many locations as they used to but folks seem to sorta want some brick and mortar options still.
I say this as someone who is required to use Claude Code at work and is generally not happy about that.
The functionality is useful enough that it will never go away. I think after the bubble pops and these companies have to actually make a profit, it will be too expensive for your average person. Businesses will want to still use it and will be able to pay what it actually costs to provide the service.
The idea that everyone is going to write emails and manage appointments and stuff like that is nuts. The value proposition isn’t there. No “home user” is going to pay $50/mo for things like that. It just won’t happen.
Why do you dislike Claude? I find that it gets the mundane aspects of coding out of the way quickly so I can focus on more complex tasks.
I dislike having to use any AI tool. Solving the problem is the fun part. Not telling the machine to solve it.
AI is here to stay. It will likely get its own ‘bubble’ crash, because it’s also a speculative bubble, but it’s not just that. Unlike NFTs, AI is a real game changer tool (whether we like or not).
The only question that remains to be decided is which country will develop the real winning ‘model’ of AI.
So far, it seems the USA have manged to keep some lead.
The EU has the brain-power to be able to become a meaningful actor too but as a EU-citizen myself I doubt the EU will ever wake up and start building anything, as the 27 of us would first need to be willing to work together and the EU itself would also need to be willing to reduce its own regulations that make it impossible to do anything, and that is as likely to happen as, say, Mr Donald J. Trump and his actually becoming a nice working to make the world a better place.
But I would not write off China: at the exact same time the USA is working hard at dumbifying itself and its population (and also making the rest of the world can only wish for the USA to not be a thing, and that is including their all-time closest allies), at that same time China is working hard at becoming a scientific and tech leading/innovating country, and they also had more than enough brain-power. Also, unlike the EU, China don’t need to learn to work together.There will be a crash / correction but everyone will continue to use Gen AI to write their assignments and make youtube videos.
No. An actual AGI will be the last invention we ever make. It’s just too valuable of an asset to not chase. Kind of like immortality in a sense that it’s hard to imagine a situation where we just collectively choose it’s not worth chasing for. Ofcourse it is and will always be.
Sure but no one actually believes that LLMs are a step on the path to AGI.
You can keep iterating on a car as much as you like and you’ll never end up with a spaceship.
Sure but no one actually believes that LLMs are a step on the path to AGI.
Dog, loads of morons think that. Hell lots of them think LLMs already are AGI.
I wouldn’t say no one but that’s besides the point anyway. I don’t feel like this discussion is about LLMs specifically but about AI broadly.
Sure but right now we’re in a Gen AI boom, not an AGI boom.
There will be new applications of AI or something entirely different. There will always be another tech boom. AI may or may not be part of it, but it’s not going away. It’s more than LLM’s.
