Depends on what is a ‘generation’ for LLMs. It would be weird to build a model which is a generation behind. My guess is that like all models, it will be considered the best until the novelty factor wears off and then it will be more or less the same like all modern LLMs - better in some domains, worse in others.
Edit: and it will probably also lead in most major benchmarks which says next to nothing about the quality.
My CTO is pushing 30k line PRs and when asked “how do you know it works” all he can say is “I’m not sure but it probably does. Our customers can QA”. Meanwhile I’m cleaning up half vibed messes from my coworkers that demo’d well.
They’re very powerful, but I think their marketing departments are even more powerful. I do wonder how many of these comments are real people.
IMO If you're not using the tools regularly, you should start from the assumption that you're being heavily influenced by the folks that are trying to sell you something. I'm using these agentic tools for writing and improving software many times a day. The way I describe this to people is that it has changed the balance between "do I need to teach the computer to do this" or "do I just do it myself" farther in realm to the former. It's really good a boilerplate which I think is a welcome thing altogether. It's not actually doing the hard thinking that's required to scale something in production, reconcile a conflict between privacy and functionality effectively for your particular customers, or determine with any reliability what will actually sell.
It will have to eventually. If the current rate of technological advancement continues, it is only a matter of time before every job will be automated. What would we do then if there is no UBI?
Die (say the exploitative elite). What is your answer to that? Don't try fighting them because they control the government and power. (Just saying it from their pov.)
As for it having permanence, I don't think it's a given, even if it's statistically likely for the majority. People have worked themselves out of poverty for ages.
> Why is everyone so worried about poison going away?
Hello, aloof galaxy-brain.
If you weren't aware, we live in a capitalist economy. In capitalism, if you don't have money you are in a state we call "poor," which means your life is difficult and your living standards bad. Most people rely on having a job to make money, and usually need a well-paying job to be comfortable and secure.
AI isn't going to change any of that. It's not going to make energy, or land or housing more abundant, in fact it will probably make all of those things more scarce.
If the future doesn't have "jobs," there will probably be a holocaust of workers before we get there. IMHO, that's more likely that some kind of utopian techno-socialism.
One farmer works full time to grow enough food to feed everyone.
One carpenter works full time to maintain housing for everyone.
The other 18 people work full time reciting incantations to the gods to prevent their destruction.
The farmer and the carpenter are too busy growing food and maintaining housing to recite incantations all day.
So each of the 18 incantation reciters spends an extra two hours every week to recite incantations for the farmer and the carpenter to prevent their destruction too.
Then a magic stone is discovered that can recite the incantations by itself, the gods are appeased, and nobody has to work full time reciting incantations any more.
How can the islanders possibly survive now that 90% of them don't have full-time jobs?
The farmer and the carpenter are no worse off, but they think it's unfair that they must continue working full-time while 90% of the residents don't work at all.
Perhaps the farmer and the carpenter should destroy the magic stone so that everyone will have full-time jobs again.
Or is there some other way the islanders can survive after 90% of the work has been eliminated?
Here's mine: the island has a leader/owner, he takes the stone for himself. He tells the 18 incanters that they're obsolete, and lectures them about how the future is glorious and its all their fault they can't change with the times. Then he allocates more of the resources to himself. Sure, he doesn't "need" more, but it's abhorrent to pay moochers who are providing no value to him. He owes them nothing. The moochers die off, slowly or quickly, it doesn't really matter.
If the islanders are ever to band together and liberate themselves by seizing the magic stone, then they must first stop praising the owner for making them work, and they must stop plotting to destroy the magic stone that could liberate them from their toil.
> If the islanders are ever to band together and liberate themselves by seizing the magic stone, then they must first stop praising the owner for making them work, and they must stop plotting to destroy the magic stone that could liberate them from their toil.
Dude, your story proves nothing. It's a fantasy. It doesn't map to the society we live in, or any realistic social change. The islanders didn't do anything before the "magic stone" to change their situation ('cause that ain't the first stone, there have been dozens before it), and they're unlikely to do anything different afterwards.
Also, if you think about your story correctly: the "magic stone" doesn't liberate the islanders, it destroys their power to make change.
They can try, but it will be very difficult when the owner has mechanised autonomous killer bots that never sleep and won’t stop until every single incanter has been wiped out.
It seems like LLMs will result in "service abundance" sooner than "material abundance." Both since progress in robotics seems to be behind that of LLMs, and because the US doesn't even manufacture most stuff anymore anyway, so we aren't self-sufficient in that sense.
We'll reach a strange situation where "service jobs" are commoditized while hard labor becomes the bottleneck. Worse, the US had transitioned to a service economy and devalued hard labor, and that also seems unlikely to change. Maybe the people going on about the permanent underclass were onto something.
With AI, you'll be living in a box under the street, but it'll be great because you'll have an AI therapist available 24x7 to talk to about your problems.
I'm pretty sure. If our society was capable of that kind of adaptation (less/no jobs), it would have done it by now. In fact, the impulse has been in the other direction. AI's not some magic thing, in this way it's no different than any other labor-eliminating technology.
> I think the future will be here faster than we imagine
Yeah, but that future ain't going to be the utopia promised by AI's salesmen. They're salesmen, and they're saying whatever they need to say to get you to buy in.
Isn't a house too personal that you'd want to get a professional architect with experience to design it, and sign off on it? Even if they used advanced tools like CAD and copy pastes 8/10 of it?
Sure, you can probably one shot notepad.exe but it has no meaning. Meaningful work isn't going anywhere, for the reason that meaningful work lives and lives on by people for people.
No one wants a vibe designed car, unless you are one of those psychos that has no tastes and doesn't care about anything.
Have you worked with a professional architect. Cost adds up fast, and you get 1-2 iterations?
I'd love to work and vibecode the house to my full liking, assuming that the agent harness will take care of all the nonfunctional things (stable design, zoning etc). Same for car if I could customize it I would.
(I definitely don't like the ramifications of it on the economy/jobs, but the above are pure consumer wins, no doubt)
> I'd love to work and vibecode the house to my full liking,
Instead of deadcode, it'll leave you with a few extra secret rooms that have no doors or windows :)
The reason you wouldn't want this is cost. The cost of building a house is marginally affected by designing it with an AI agent. Most of the cost is bricks, etc -- material.
I absolutely would want a vibe designed house or car. Vibe to completion, then have a now higher paid expert architect/engineer review the plan to confirm it is safe to proceed.
Until we can reliably trust these models to skip that last step, there is nothing inherently wrong with using AI to create, so long as it is manually reviewed for accuracy at the end.
The percentage of people employed in agriculture dropped from 80% to 2%. The market will be full of people who are willing to learn everything about AI in order to have a comfortable and highly paid job.
Becoming an electrician would be a downgrade or even impossible for some people.
For the record, I think AI replacing highly paid "sitting behind a computer" jobs would be good for the society, but probably not for most people having these jobs.
Being ahead of Google is less about raw model quality and more about shipping usable products fast. Anthropic’s advantage seems organizational as much as technical. If Sonnet 5 really halves inference cost while improving reasoning, that’s more disruptive than any benchmark win.
I keep trying to use Codex CLI but I love using claude --dangerously-skip-permissions but this seems impossible to do in codex, and it just asks me to approve every command per session. Am I taking crazy pills or is there a way to make codex just run in yolo mode?
Edit: and it will probably also lead in most major benchmarks which says next to nothing about the quality.