nisegami

>There's no amount of "nicer to use" that can replace security, at least certainly when selling to enterprises

We can lower our standards instead

>Do you use qwen for any coding?

No.

>Do you make a frontier model orchestrate it?

Also no.

>Are you running with llama.cpp?

Yes, on Arch Linux.

>Sorry for the barrage of questions

No worries.

My Qwen testing has been with Hermes trying to do mundane (to me) things that hosted models understandably refuse to do, even open-weight ones like Deepseek V4.1 Flash. Example: I tried to get it to find a torrent for $tvshow and add it to qbittorrent (connected to hermes via MCP already). It certainly tried and I think frontier models would have done better. My biggest issue was context space more than anything else. I had to drop down a quant level to increase it to 128k. On 64k it was constantly compacting.

There's this weird tension where small models are getting a lot better, but the gap between the frontier and "what I can run on my gaming GPU" at home isn't necessarily getting smaller. That said, Qwen-3.8-27B-Q3 runs on my 9700 XT and performs well enough for tool calling in agents, which is a critical floor in local model performance imo.

Yes, I like the phrase "you can't expect someone to put the time and effort in to read/etc what you didn't put the time or effort in to write/etc".

Alas, it will not. It will instead dramatically increase suffering in the world while leaving most humans alive.

I think the history of the FBI suggests this will continue long after Trump himself leaves office, effectively becoming a state run gang.

Adding complexity would imply you're doing both. In this case it would be more correct to say you're trading the complexity of key distribution and access control via keys for the complexity introduced by tailscale ssh.

Is there any reason not to take the buyout?

>They could network their smart glasses and leverage their ubiquity to help people find lost pets — but strictly only lost pets, not people — but then Something Labs might ship a tool to make a person's face look pet-like enough to turn Find My Pet into Find My Ex.

This exists btw, from Ring: https://ring.com/ca/en/search-party

At least Gemini is an okay model independent of how Google is leveraging for consumer experiences.

>Google, Apple, and Microsoft are each in a stronger position to win consumer AI because they're guardians over the foundations of the user's experience.

From everything I've seen, they're all too large and dysfunctional to make a competent play in consumer AI. Anyone in there trying to work on it is too far removed from end product and is answering to too many masters.

>But what I wonder: can we legitimately grind on Physics or curing cancer?

If we can enable a feedback loop, yes absolutely. But feedback loops for things in the physical world like these are measured in months per cycle usually.

There are really fringe perspectives out there that imagine capitalism as a living entity of which we individual humans are simply parts of the whole. Governments that represent the economy over the people are exactly what you'd expect if that were true.

>"Information wants to be free".

I agree wholeheartedly and in keeping with that, I call upon frontier AI labs to release both their weights and training sets.

You're right, but the scandal is so upsetting that I personally cannot ever look at Fujitsu the same way even. Also, I think you're neglecting that a large part of why it became so bad was that Fujitsu claimed there was no issue with their software. Even if it were built by a company they acquired, the response to the developing situation was something that the parent company should have stepped in to oversee.

Things are going to be alright.

The golden path I dream of has ASI replacing human-run governments

Non-US based producers too, I can already see the difference in my country's economic position.

Ditto. I loved my Airpods Pro 2 but it was really clear Apple did not want me as a customer of their headphones. In a practical sense, the recent open source work would have mitigated most (all if it supports software updates) of my issues, but I still won't be buying any more Apple audio products out of principle.

One question has been nagging me for this situation. Levent Alpoge works at Anthropic and would presumably have some knowledge of "how the sausage is made" and I would hope he would be aware that his collaborator was utilizing LLMs in some capacity for their joint work. Would he not have guided him otherwise if it were an open secret that this kind of thing was a possibility?

Representing a 3D sphere-like surface on a 2d plane definitionally requires compromise.

>I build the base manually and then I can iterate with AI faster once the fundamentals are solid.

I've had tremendous success with this approach as well.

Isn't that essentially a manual passkey?

>Chasing down a link in your email is much slower than having your password manager fill in the long unique random password and hitting "log in".

That's basically a passkey without its special API.

How does capital even produce anything?

>How do you prevent credential stuffing attacks?

Passkeys or magic links seem like the way forward here.

>Those spaces should be subject to some form of public inspection but there's probably a good argument to not allow random cameras in.

Earlier this year various elected officials tried to perform exactly that kind of public inspection of ICE facilities and met a lot of resistance. I know that some subset of these were just publicity stunts, but despite that, the fact remains that the public needs to be aware of what detention in those facilities is like and if the DHS makes choices that result in random cameras being the only way to do that, then they should reap what they sow.

There's a lot of similar projects out there that are closer to what you're looking for. For example, I used nanoclaw for a bit before switching to Hermes.

>So a "hello world" program is now political?

There's an implicit political statement in being able to write a simple one line hello world program and run it, using free software and information. Consider the counterfactual world where to even begin programming you have to pay for certification and training or specialized equipment/software. I know paid compilers, proprietary languages and closed specifications exist, but we can still just save print('hello world') to a file and run it using python which is freely available, or even better we can open devtools in our browser and jump into the console and run console.log('hello world') without even installing anything.

Isn't the obvious response that they should offer a comparable comp package without any sort of stock aspect?