> 1. What's the bottleneck in ramping up RAM production? Is it the availability of silicon itself? Or the factories are at capacity?
For a RAM manufacturer, the incentive is to ramp up production AND prices. I doubt any of the names in the business is doing any work at all to lower their unit prices.
The current pricing isn't sustainable, and if you try to wring out the maximum amount of profit, you're gonna have competition spring up. The Chinese are probably salivating at this opportunity. Previously they would've had to sell memory at discount rates to get anyone to switch over to them, but now they'd have customers lined up for years if they were selling at the prices Samsung/Micron/SK Hynix were selling a year ago.
> I was signing up for Hetzner years ago and it asked me to upload my passport to use their service.
I don't really understand what bothers you so much about providing a photo of a "passport" (if you are an European citizen they require a ID card) but credit card info didn't registered as a concern worth noting. Can you explain what is the difference?
Credit card is a largely fixed risk of financial loss, with some legal safeguards for recovery, and the ability to get a replacement card with a different number. Passport carries an open long-term risk of impersonation and you can't just get a new passport because some company has a copy. Just the financial side of that risk can have much greater impact. Unless a company has a legal requirement to "know your customer", e.g. a financial institution, this is a red flag.
Couldn’t have put it better myself. Even with payment processors, most they ask for is SSN and business EIN.
When I read about the WireCard scandal, the KYC stuff sent to them over the years is probably in the hands of foreign intelligence already. That’s what gave me pause.
When many sites are collecting these photos, it increases possibility of them leaking. Since these are also used for KYC process in crypto sites etc, this in turn increases risk of identity theft.
I'm a Hetzner user in the US, but I pay for it with PayPal and was never asked to give my passport or identity. Americans are very rarely asked for these documents online, and even then it's typically only for government or financial services. It's also drilled into us that this info can be used for identity theft, so it's only natural to be wary of any non-government entity asking for them.
FWIW, if Hetzner had asked for my passport when I signed up, I would not have given it either.
> The whole 'charity' persona he put on afterwards is just PR.
Indeed. Even in the middle ages rich people leaned heavily on charity to whitewash their legacy. I mean, the Catholic church even made this accessible to the masses through indulgence.
> Whenever something gets subsidized by the government it inevitably becomes dominated by exploitive petty politics.
Your US-blend of anti-state brainwashing is showing. There is nothing inherently different in the for-profit status of an organization that prevents the occurrence of "exploitive petty politics". You see those from any organization from homeowners organization to full blown FANGs. I mean, have you ever paid attention to the crap being pushed by the likes of Tesla/SpaceX/Twitter?
> Then the ideologues and political commissars showed up
I think you're seeing the world through rose-tinted glasses. In some FLOSS circles the discussions were dominated by ideologues, to the point some discussions seemed like Monty Python skits. I mean, your choice of window manager, let alone Linux distro, was something you'd be judged by.
'Were'? There are still people claiming that fronted isn't real coding. It is deeply embedded in human nature to define an us v them, and if skin colour is off the table substitutes must be found.
> This seems to be an administration that is dead set on antagonizing every single part of the world. Either enemy or ally.
Let's not forget the Trump administration threatened two separate NATO allies with invasion and annexation, and then had the gall to complain they were not helping them attack Iran.
> I am saying this probably is "silly behavior by a government" and it is a milestone that points towards what the future may look like. Why can't it be both?
Here is why it's unlikely this is anything other than "silly behavior by a government":
- some benchmarks show GPT-5.5, Gemini 3.1, and even Claude Opus outperforming Claude Fable, and yet it's Fable which is restricted.
- some benchmarks still show the likes of Kimi 2.5 outperforming any Claude model, and DeepSeek is getting equivalent scores (a few tenths of a percent difference)
> Do you think that Chinese labs will continue to release open models forever (...)
That's immaterial to the discussion. Even if China forced Chinese labs to restrict access to all models, the truth of the matter is that Trump's administration to restrict access to US-based models does not prevent others from having access to models that are as capable or even better.
You’re completely overrating these benchmarks and it’s landing you at a nonsense opinion. Just actually use the models and you will see that the gap is significant.
It should be easy for a company like Anthropic to prove this beyond a doubt. Why don't they? Why don't they have a collection of prompts and side-by-side comparisons with other models showing how far ahead they are?
I think it's mainly because the difference in models at the frontier isn't "response to prompt X", but rather "coherence with 500K tokens of context and instructions in play"
All that says is some benchmarks aren’t worth the tokens it takes to evaluate them. Mythos is clearly capable of finding zero days other models can’t, and Fable is close enough to be lumped with it.
> Mythos is clearly capable of finding zero days other models can’t
I'm unconvinced that this is anything more than proof of work and marginal improvement that other models will catch up with, perhaps as early as to next week. Lots of other current-gen models will find vulns that can be chained together if you're willing to burn enough tokens on the task, and Fable is an absolute token incinerator.
> By building pain into the system. If your hands dealt with injury directly without sending pain signals up to your brain, you'd never change the behaviour that led to that harm or reconsider your priorities.
I don't think that's it. Emergent problems require attention and action from leadership, who in turn can make the problem visible to higher ups. This creates signal, and positive feedback when the problem is fixed or mitigated.
If the problem doesn't exist to begin with, there is no signal. Managers don't get to show their fast-acting skills, and there are no heroics to speak of.
So ultimately poorly maintained and managed projects who deliver fixes for problems of their own doing create a perverse incentive, whereas no one is lauded or promoted for doing normal day-to-day things.
Well I think it is even more complex. If you're a plumber in a rotten system of pipes the whole company depends on, you can fix issues day in and day out, without speaking a word and they will notice everything is a bit unreliable and thus you do a bad job. You could do the exact same work, but make a big thing about every major fix, warn people a week ahead, give them the feeling the company depends on it and then do the exact same work and tell them how you fixed it. Suddenly you did a good job, despite you literally doing the exact same thing with your hands.
The difference is how it was communicated. Most non-Tech/non-infrastructure-people got no clue about these things. If they know you're battling the demons of plumbing on their behalf they will thank you, if you're the weird guy that has smeared dirt in the face and is seen once a week while the plumbing fails ever so often, guess what.
That means even if the problems and their fixes remain the same, the communication around them really matters. Tech people can be extremely bad with this. And if we're talking IT it is really the plumbing that holds the company together.
> IMO comparing different models is like comparing songs or paintings or modern art.
I don't think this is that subjective or vague.
There are a couple of crisp metrics that can be used to evaluate a model:
- given a prompt, does it finish a task (times X tasks)
- how much did it cost to finish the task
- how long did it took?
If all models are able to handle a class of tasks, they perform equally well.
If a model costs much more to finish a task, it is worse than other models.
If a model takes longer to finish a task, it is worse than other models.
The ugly truth is that since the GPT4.1 days, new model releases have shown diminished returns. Context windows were increased, reasoning steps help improve the usefulness of a user's prompt,... That's it. Even those are UX improvements, instead of huge breakthroughs.
"Diminishing returns", so are you claiming unironically that GPT4.1 can achieve anything Fable 5 can?
Or just that it's so much cheaper that the cost/benefit ratio is better?
Also "finish a task" is also subjective. I can "finish the task" of building a table, but it will be a shitty table. Are you also measuring the quality of the result - which is subjective again?
> "Diminishing returns", so are you claiming unironically that GPT4.1 can achieve anything Fable 5 can?
I see you felt compelled to use the weasel word "anything" to put together an argument. That suggests you are very well aware that the difference between older models and the latest and greatest is not that significant, as you need to resort to coming up with a single example, any example at all no matter how far fetched, to try to put together a case.
And that says it all.
> Or just that it's so much cheaper that the cost/benefit ratio is better?
That too is another definition of quality, isn't it?
If you have two tools and one does the same job but is both cheaper and faster, it means it it objectively better.
> Also "finish a task" is also subjective.
No, it isn't. If you supply a prompt and you have a definition of done, and a model executes it and delivers what you asked then it finished the task successfully.
> I can "finish the task" of building a table, but it will be a shitty table. Are you also measuring the quality of the result - which is subjective again?
Nonsense. If you feel the need to put up strawmen then it's up to you to justify them. Please define "quality" and prove that a model such as fable has such a radically different output that in comparison the output of older models is "shitty".
I understand you feel the need to keep the hype bus going, but you need more than strawmen, weasel words, and hand waving to keep that hype afloat.
And the truth if the matter is that the models introduced in the oast year don't introduce any breakthrough and struggle to show significant improvements over older models.
> Clone simonw/micropython-wasm from GitHub and research how this could use a full Python as opposed to MicroPython
I might be missing something important but that doesn't seem to be an impressive task.
On a surface level it sounds like the taks requires gathering calls to MicroPython-specific libs, assess which ones are not compatible with Python, and proceed to determine how to replace the ones that are incompatible.
From that first iteration, the rest would boil down to troubleshooting the issues missed on the first shot.
I would be extremely surprised if the likes of GPT4.1 wasn't already capable of handling that task.
So, beyond Claude Fable finishing a task, what exactly is the differentiating factor?
For a RAM manufacturer, the incentive is to ramp up production AND prices. I doubt any of the names in the business is doing any work at all to lower their unit prices.
reply