p_stuart82

5 ms·

31.
by p_stuart82 5mo ago
$250b committed to azure helps. especially when some of that is your own investment coming back.
32.
by p_stuart82 5mo ago
the thing is it doesn't even feel like mortgaging. shipping, features going out, everything looks fine. then something breaks and you realize you can't debug your own code without asking the model again.
33.
by p_stuart82 5mo ago
basically discount Kaggle. still get people poking at it, just none of the writeups or who-gets-paid drama.
34.
by p_stuart82 6mo ago
not just the cache though. every time you stop and come back, it basically reloads the whole session. if you just let it keep going, it counts like one smooth run. you hit the wall faster for actually checking its work.
35.
by p_stuart82 6mo ago
tbh ~1-3% PPL hit from Q4_K_M stopped being the bottleneck a while ago. the bottleneck is the 48 hours of guessing llama.cpp flags and chat template bugs before the ecosystem catches up. you are doing unpaid QA.
36.
by p_stuart82 6mo ago
for software engineering? not because of the typing. the signal is every time a human has to grab the wheel. that's a label for what the agent still misses.
37.
by p_stuart82 6mo ago
exactly people paid the premium so somebody else's OAuth screwup wouldn't become their Sunday. and here we are.
38.
by p_stuart82 6mo ago
IMO nobody was paying for magic compute. they're paying to not touch ten years of glue. if agents eat that glue, the moat gets thin fast.
39.
by p_stuart82 6mo ago
IMO it doesn't flatten design into one thing. it splits it. cheap obvious work at scale, and a way smaller premium tier for real authorship. the middle is what actually gets crushed.
40.
by p_stuart82 6mo ago
the awkward part isn't just about reading sensitive files. search, listings, direct reads, browser and computer use all sit behind different boundaries. hard to tell what any given approval actually buys or exposes.
41.
by p_stuart82 6mo ago
caveman stops being a style tool and starts being self-defense. once prompt comes in up to 1.35x fatter, they've basically moved visibility and control entirely into their black box.
42.
by p_stuart82 6mo ago
yeah they took "i pick the budget" and turned it into "trust us".
43.
by p_stuart82 6mo ago
you're welcome.
44.
by p_stuart82 6mo ago
separating codebase and leaving 'cal.diy' for hobbyists is pretty much the classic open-core path. the community phase is over and they need to protect their enterprise revenue. blaming AI scanners is just really convenient PR cov
45.
by p_stuart82 6mo ago
yeah the desktop app forgets it's the desktop app. claude code feels local right up until the api starts coughing up 500s. same thing, just in a terminal instead of a window.
46.
by p_stuart82 6mo ago
locally? sure. stacked changes in jj are great. but the moment you push to GitHub, the review UI still thinks in SHAs. a lot of the pain just moves from the author to the reviewer.
47.
by p_stuart82 6mo ago
gave it "no instructions" but gave it memory files, a twitter account that pings it back, and hacker news. that is the instruction.
48.
by p_stuart82 6mo ago
defaulting to strip location on share, fine. demoting plain old <input type=file> into "find a usb cable" / "go build an app" is a hell of a line to draw
49.
by p_stuart82 6mo ago
yeah 400 kbps is almost the easy part. you still need a line, a handset, and apps that still run on the cheapest phone around. hard to call that universal in practice.
50.
by p_stuart82 6mo ago
performative is right. files & folders says blocked. open panel access still works. the pane only knows about one path
51.
by p_stuart82 6mo ago
blunt mode isn't the missing feature here. reading the source is. it literally says "this entire paper is made up"
52.
by p_stuart82 6mo ago
19k tokens per session and the skill triggers don't even check project scope. you're paying that overhead on every non-vercel repo
53.
by p_stuart82 6mo ago
$2-4/hr always sounds cheap until you multiply by wall clock and reruns
54.
by p_stuart82 6mo ago
late isn't the problem. private preview api and no reason to switch. that's just another hosted model
55.
by p_stuart82 6mo ago
models are improving. the pricing already assumes they're ready for prod. that's where the fires start
56.
by p_stuart82 6mo ago
subscription leftovers are noise. the real switching cost is the harness glue. prompts. tool calling quirks. evals. auth. retries. all the weird failure modes your team already paid to learn.
57.
by p_stuart82 6mo ago
lol if your data sovereignty still depends on someone else's control plane, it isn't sovereignty. same landlord. lower in the stack
58.
by p_stuart82 6mo ago
it feels like the modern recurrence of the early 2010s bootstrap templates. we figured out how to automate building sites instantly, but at the cost of making the entire web look exactly the same.
59.
by p_stuart82 6mo ago
the irony of modern software engineering: we spent decades perfecting deterministic algorithms, and now we're basically just shaking a black box and hoping the magic rocks align.
60.
by p_stuart82 6mo ago
the hardware ROI is insane right now tbh. a $200/mo sub is literally paying off a second gpu in less than a year.

More ›