1,211 karma · joined October 12, 2021
I find it odd you see otherwise.
And VS 2026 is faster than VS Code these days, now that they spent time optimizing it.
Exposure breeds familiarity, which breeds desire, I think.
Even if all you hear is an occasional ad about the company, you might get curious and research them yourself.
The VPN side - it feels more like a honeypot than anything else to me.
I'll sometimes use the VPN mode from Proton, but I wouldn't sign up for just a vpn - I simply assume they are selling my data the same as everyone else. The only way to convince me otherwise is an open-source non-profit vpn properly overnighted by the community.
They'll be anti data centers until they are no longer the opposition party. Then suddenly both Republicans and Democrats will change while MAGA/America First and woke 2.0/socialists sides won't have changed at all.
Isn't the political party system great!
The exact same ones are now pushing $300/ssd and Samsung's (which was $300 18 months ago) is now $500.
The prices are insane.
That probably is the cheapest per 1 GB vram right now. However, the R9700 a few months ago was only $1350/32GB or $42/1GB VRAM. And everyone knew it, because the card is now $1800 or so with supply issues.
I think the market discovered the reasonable min is 32GB. That runs Qwen3.8 Q6 with 100k+ context all in vram. 24GB does not.
Seems like simplify or just erradicting the tax code would be more just at this point.
My stack uses a local llamacpp ROCM build, modified by https://github.com/stew675/llama-cpp-rdna-boosts
The build uses my machine's ROCM and I just link it instead of embedded all that in an image, like the published ones do.
Then deploying it via docker compose to my local machine. The models I want to run are stored in a docker volume, so it's easy to switch the model while keeping the same build.
The biggest negative for me is that I don't think it's enough to do parallelization through llamacpp (I'd prefer to work on one thing faster than two things slowly). I haven't gotten vllm working yet. Also, once you have one, you constantly start thinking about a second card...
However, that also means AI will be even more lobotomized since the cloud providers will panic over being liable
Sadly, the r9700 has gone up in price, but it's still the cheapest route. I use It on an old x99 Linux desktop and it achieves 50 tps decode and 1000+ tps prefill.
Why trust Google of they aren't acting in your interest?
Now, moving a steam account into a trust might be a breach, if they care. Even using a trust to buy a game might also be a violation of some form, perhaps, buried in the toc.
But the trust itself is a legal entity and it can own things. Just because the people behind the trust changes doesn't mean it was a transfer on the public side of the trust.
No one gets pulled over for turning light violations for safety reasons, but that's beside the point. I just hate how many laws we have that are just selectively enforced.
I force all work to be split up into chunks, properly investigated, spiked, and planned out, and then implemented via TDD (with only system-boundary mocking permitted). It's not perfect, but I've seen regression bugs very much cut so far.
Beyond that, if it's a GUI app, I also require the TDD cycle to require GUI tests with screenshot evidence, and proper assert statements.
In other words, my email can be signed in from anywhere on purpose. It's a feature. The chances of me forgetting my password is zero, because I can back it up 7 way to Sunday.
A passkey seems to miss these basic accessibility issues. For a random website, that's fine, but for the service all of websites use to auth me, access is more important than locking it down to one device.
One domain, one account.
They don't really have a leg to stand on.
You can't nationalize another country's business, especially when their assets are floating in space disconnectes entirely from your country.
If they want to regulate it, they need to build it first. Ukraine has no ability to regulate Starlink. It's not theirs to regulate.
Few people are thrown into a codebase that they understand because they wrote it. Not every developer is issued their own greenfield project and said go have fun.
Any programmer should be able to hack one up. LLMs are good at _programming_, but I haven't seen evidence they are good at engineering yet. But then again, many people with the title "software engineer" are not great at it either.
However, it doesn't take away the concern of architecture, design, etc. and questioning if the current solution is well designed or not.
In other words, if you can't question the AI and refute it in a topic, you aren't expert enough to use AI in that area in a engineering manner.
Granted, not everything needs an engineer behind it. A shack to store some tools will survive long enough probably
The (classical) Latin alphabet can be fully described by the English alphabet.
The government should face a punishment for targeting constant inflation.
It's quite uncivilized to see a situation where someone is mugging someone and hundreds of people are just passing it by like it's another Tuesday