It has always baffled me how quickly, and how voraciously, people started to rely on privately owned AI systems.
AI is not something discovered by scientists and plucked out of the ether. It's engineered and controlled, for profit, by corporations which have demographics and KPIs. These companies don't owe you anything, and they make no promises.
If you're running a business that deeply relies on AI, you might as well add Sam Altman to your board of directors--because he has just as much control over your company as you do. If they have a bad quarter and need to increase rates by 1000%, your choices are to pay up or shut down.
This Mythos situation is just the beginning. Not only do they have everyone hooked, but they've actively stalled the personal skill growth of millions of people who fell into vibe-coding rather than genuinely learning. And now they have that choice: Pay up, or shut down.
The same corporations that insist upon private Maven repositories to control all code dependencies are nevertheless fine with establishing a massive dependency on a privately-held corporation in order to write software that hardly anyone in the organization understands. When I really think about this and how it plays out in the long run, I feel like I’m taking crazy pills.
I can't run my business without electricity. Yet we don't fear of its access being revoked. Sam makes the comparison of intelligence to electricity a lot. So we are on the path to these systems becoming utilities.
Whatever is in Mythos will be open source in 6mos-1yr tops. You might not have the GPUs but you won't be locked out of the capability.
We're still not at the point where one person with a coding agent can max out their salary in effectively using credits, so the capability is still well within reach of the vast base of the industry.
Meaning that for now, most people who want to pay for the product (which IMO is pretty reasonably priced for what it does) will be able to get the product.
The economics will make sure of that. The market is ripe for someone basically copying the likes of Mythos and pricing it competitively.
That assumes “more model” is the part that differentiates successful ideas from unsuccessful ones.
Governments and corporations controlled enormous mainframes far beyond the compute available to the hacker kid we were waxing nostalgic about, didn’t they? Not to mention the PhDs, the mountains of capital, and so on?
I think that remains to be proven. The context was 16-year olds being able to freely build things. They still can do that as before. Not everything is a competition.
>> We saw yesterday that expert orchestration around small, publicly available models can produce results on the level of the unreleased model.
This is false. Yesterday's article did not actually show this, and there are many comments in the discussion from actual security people (like tptacek) pointing that out.
It has always baffled me how quickly, and how voraciously, people started to rely on privately owned AI systems.
AI is not something discovered by scientists and plucked out of the ether. It's engineered and controlled, for profit, by corporations which have demographics and KPIs. These companies don't owe you anything, and they make no promises.
If you're running a business that deeply relies on AI, you might as well add Sam Altman to your board of directors--because he has just as much control over your company as you do. If they have a bad quarter and need to increase rates by 1000%, your choices are to pay up or shut down.
This Mythos situation is just the beginning. Not only do they have everyone hooked, but they've actively stalled the personal skill growth of millions of people who fell into vibe-coding rather than genuinely learning. And now they have that choice: Pay up, or shut down.
The same corporations that insist upon private Maven repositories to control all code dependencies are nevertheless fine with establishing a massive dependency on a privately-held corporation in order to write software that hardly anyone in the organization understands. When I really think about this and how it plays out in the long run, I feel like I’m taking crazy pills.
your choices are to pay up or shut down.
Another choice is to switch to a different model, perhaps open source this time.
Package manager incidents (like leftpad) have shown that just because it's open source doesn't mean it can't do damage to your project.
Another choice is to write code and learn. Especially if you are 16 and have all day.
1 reply →
I can't run my business without electricity. Yet we don't fear of its access being revoked. Sam makes the comparison of intelligence to electricity a lot. So we are on the path to these systems becoming utilities.
Electricity is heavily regulated. Is there any evidence that LLMs will be the same?
2 replies →
I would bet any amount that when the time comes to turn AI into a utility, they will fight it tooth and nail.
Whatever is in Mythos will be open source in 6mos-1yr tops. You might not have the GPUs but you won't be locked out of the capability.
We're still not at the point where one person with a coding agent can max out their salary in effectively using credits, so the capability is still well within reach of the vast base of the industry.
Meaning that for now, most people who want to pay for the product (which IMO is pretty reasonably priced for what it does) will be able to get the product.
The economics will make sure of that. The market is ripe for someone basically copying the likes of Mythos and pricing it competitively.
That assumes “more model” is the part that differentiates successful ideas from unsuccessful ones.
Governments and corporations controlled enormous mainframes far beyond the compute available to the hacker kid we were waxing nostalgic about, didn’t they? Not to mention the PhDs, the mountains of capital, and so on?
My money’s on team human.
You can do things without AI. That hasn’t really changed.
The point is you won’t be able to compete with just your brain
I think that remains to be proven. The context was 16-year olds being able to freely build things. They still can do that as before. Not everything is a competition.
We saw yesterday that expert orchestration around small, publicly available models can produce results on the level of the unreleased model.
I take a contra view and instead see this as fuel on the fire for tinkering to squeeze advanced functionality out of more available things.
It has always been like this, the amateur improvising tooling and equipment to outdo companies with comparably infinite resources.
>> We saw yesterday that expert orchestration around small, publicly available models can produce results on the level of the unreleased model.
This is false. Yesterday's article did not actually show this, and there are many comments in the discussion from actual security people (like tptacek) pointing that out.
From what I can tell, this was not clearly settled.
Your example author, actually corrected themselves saying LLMs “possibly” could perform successfully: https://news.ycombinator.com/item?id=47732696
1 reply →
You can’t use your own brain?
The point is you won’t be able to compete with just your brain
I’m all about AI and not-AI… but the question is, can you use your own brain 24 hours a day?
I don’t agree with vibe coding, I see the appeal of an AI ticking through my code at night to see if tests could be better or I missed something, etc.