baq

26 ms·

751.
by baq 7mo ago
> Because to hell with UX when it comes to security. I don’t think you have any idea how wrong you are.
752.
by baq 7mo ago
macs are amazing, you can close the lid on a macbook and the computer will be mostly asleep and won't burn down the house; and the best part is it'll work when you open it again. amazing tech. macOS though, that's getting wor
753.
by baq 7mo ago
the swap and vm tuning required on Linux to make it work even somewhat reasonably compared to Windows or macOS is batshit insane - or brain dead to quote Linus. you shouldn't need to do any of it and yet if you don't you risk hard
754.
by baq 7mo ago
There's a nonzero chance military intelligence agencies of multiple countries know exactly where that plane fell, but none can say anything, because that would reveal the true extent of their capabilities.
755.
by baq 7mo ago
my favorite kind of hn comment. > The link to the homepage is literally on top of the README. the 'technically correct' part > You're just being a complainer for complaining sake. the 'completely missing the point&
756.
by baq 7mo ago
You’ve listed a whole gamut of reasons why having the process is essential, so they weren’t wrong ;)
757.
by baq 7mo ago
A webui for creating vms is certainly convenient if you’re doing it once a year like me.
758.
by baq 7mo ago
Yes. One image won’t kill you. Or just refer an external static image.
759.
by baq 7mo ago
Agreed! I’m using cockpit on my home assistant box and I wasn’t even sure this is the same project since I log in with muscle memory.
760.
by baq 7mo ago
Discord webhooks aren’t too bad… but the proper bot thing is ridiculous. They really lack a development mode server, having to know everything about oauth and token permissions before even starting is bonkers and why do I even need an app i
761.
by baq 7mo ago
Nobody wants to sacrifice growth today for stability tomorrow and there are good reasons why this stance makes sense. That said I’m all for it, too bad the supply chain disruption that this mess will cause will make it twice as hard as it c
762.
by baq 7mo ago
Funding is as good as gone until the Iran mess is over.
763.
by baq 7mo ago
Favelas are a local optimum which systemically is very difficult to get out of and is a sign that the regulator is powerless. It's a great example of a market failure.
764.
by baq 7mo ago
But zoning is required to maintain order. Nobody wants anybody to live in favelas. As with everything the regulator needs to strike a balance to make the market work.
765.
by baq 7mo ago
LLMs in teacher mode instead of solver mode can be great. ("review this change" is kinda sorta teacher mode.)
766.
by baq 7mo ago
Given how the previous conspiracy theories turned out we really needed some unconfirmed ones, thanks.
767.
by baq 8mo ago
Not al all, it’s what vs how, completely different beasts. How is for engineers to solve, but the business sells the what.
768.
by baq 8mo ago
Yes? If you’re an auto factory, you might care, but an auto maker cares about minimizing cost and maximizing revenue within the regulatory constraints. Nowhere is there a requirement to care about how the car is built, there are requirement
769.
by baq 8mo ago
A bit off topic, but reading your post I suddenly realized that if I read it three years ago I’d assume you’re either insane or joking. The world moved fast looking back.
770.
by baq 8mo ago
it isn't gospel, it's perspective. if you care about the code, it's obviously bonkers. if you care about the product... code doesn't matter - it's just a means to an end. there's an intersection of both views i
771.
by baq 8mo ago
> So how should you use an LLM to contribute? > Use an LLM to develop your comprehension. Then communicate the best you can in your own words, then use an LLM to tweak that language. If you’re struggling to convey your ideas with some
772.
by baq 8mo ago
it's only fraud if somebody finds out
773.
by baq 8mo ago
yesterday I had to tell a frontier model to translate my code to tla+ to find a tricky cache invalidation bug which nothing could find - gpt 5.4, gemini 3.1, opus 4.6 all failed. translation took maybe 5 mins, the bug was found in seconds,
774.
by baq 8mo ago
they will, but the jagged frontier is fractal and each one will have different capabilities; you'll want to mix models and to get best results consistently you'll need to.
775.
by baq 8mo ago
proxmox itself should work, but the installer has a much older kernel and didn't have the network interface driver.
776.
by baq 8mo ago
an intel n100/150 minipc is perfect, if you get an old enough build (don't ask) you can install proxmox on it and set up hass as a vm as God intended. if you have to ask, mine was too new and had to settle for a non-LTS Ubuntu Ser
777.
by baq 8mo ago
disagreeing with this statement is admitting it's gambling and not a market.
778.
by baq 8mo ago
true - but the way LLMs are trained, google's RLVR is different from anthropic's is different from openai's. you'll get very good results sending the same 'review this change' prompt (literally) to different mo
779.
by baq 8mo ago
it's the easiest as it's ever been to get started in a foreign code base: start up the agent and ask questions. more or less instant answers, basically zero confabulations nowadays. ...but since it's so easy to deliver stuff
780.
by baq 8mo ago
typical experience when only using one foundational model TBH. results are much better if you let different models review each other. the bottleneck now is testing. that isn't going away anytime soon, it'll get much worse for a bi

More ›