HNHacker News
TopNewBestAskShowJobs

mirsadm

1,302 karma · joined February 12, 2012

email: mirsadm@gmail.com
submissionscomments
mirsadm··on Apple M6 Pro Achieves the Highest Single-Core CPU Score in Geekbench 7
If anything it's the exact opposite. They're ridiculously powerful with more memory bandwidth than pretty much any equivalent PC. It is easy to make sloppy software because you can get away with so much.
mirsadm··on More questions about whether researchers can trust OpenAI with unpublished math
They consume everybody's hard work then sell it to all competitors. What a deal.
mirsadm··on There's no such thing as a small software team anymore
And they're really good at adding more code to reduce complexity. In particular Claude is absolutely bonkers at the moment. I am having trouble taking this seriously especially because you mentioned Claude. Have you looked at anything it's done?
mirsadm··on GitHub Actions and Pages are experiencing degraded availability
Dunno that's my experience with most businesses. I doubt AWS is that much better under the hood.
mirsadm··on SwiftUI After 7 Years
That's just not true. I am still finding the UI taking 40-50% of CPU time with many hours of wasted time on optimising it and ugly hacks. There's no convincing some people though. Because they can always say you just don't know how to use it.

Well I've been doing this for a pretty long time and I'd like to think I'm pretty good at it.

mirsadm··on $100 AI Music Video: Claude Fable 5 vs. GPT-5.6 Sol
Yes but they would have had to spend years training/practicing which apparently is not a thing we value anymore.
mirsadm··on Android Developer Verification: Threat masquerading as protection
An older iPhone is already better than most new cheaper Android phones.
mirsadm··on The best response to AI slop and online noise is from Robin Williams
You're right to push back...
mirsadm··on U.S. government will decide who gets to use GPT-5.6
Does it? That's not even well established yet. These things aren't that good.
mirsadm··on Thomann takes legal action against Fender
I'm looking at buying one now and as far as I can tell Fender seems to make shoddy quality instruments these days. I see a lot of recommendations for PRS and others.
mirsadm··on Will It Mythos?
To me it feels like they're basically tweaking these things around the edges. I'm not seeing any difference in capability just preference. This has been the case for a while.
mirsadm··on Linux eliminates the strncpy API after six years of work, 360 patches
I'd be curious to see how much CPU time is wasted on looking for a null every time strlen is called. The extra length integer is probably insignificant compared to that.
mirsadm··on Electric motors with no rare earths
What does older mean in this context? Because some people still think the year 1996 wasn't that long ago. Modern Renault cars are fine and reliable enough. I've had 4 in my life time and had zero issues myself. I see a ton of them here in the UK and, again, they're fine.
mirsadm··on Claude Fable 5
It's the real deal. Before Fable nothing I tried worked. It has finally helped me finish my teleportation device. I can't show you or anyone the proof but trust me it's true.
mirsadm··on If Claude Fable stops helping you, you'll never know
What's 12 months lead time worth? Not much from what I can tell. Contrary to what these AI companies might tell you, if an AI model can't do it, a human can still do the work.
mirsadm··on Where is the AI jobs crisis?
This seems pretty unlikely. If it turns out to be true then you don't need a junior or senior dev you can just get a random person from the street and they could do the job.
mirsadm··on OpenCV 5 Is Here: The Biggest Leap in Years for Computer Vision
That is a very uninformed view. Real time CV is not going to be doing that anytime soon.
mirsadm··on The Structural Barriers to AI Lawyers
You can apply this same argument to everything. Code is deterministic but what is being made is often not because people don't know what they want to make. Society can choose just to make everything boring and deterministic so that computers can do everything.
mirsadm··on Google I/O
It's gotta be booking a flight or restaurant or something. They demo this every year and it never works.
mirsadm··on We tested super-resolution pre-filter for LPR OCR. It did nothing
Have you tried classic CV SR algorithms (eg SR from multiple nearby frames)? Single frame SR seems unlikely to be useful.
mirsadm··on DeepSeek 4 Flash local inference engine for Metal
I've built something like this. One issue is that LLMs are actually terrible at writing good shaders. I've spent way too much time trying to get them not to be so awful at it.
mirsadm··on Pro Max 5x quota exhausted in 1.5 hours despite moderate usage
The way Claude/Codex behave is entirely consistent with how every vibe coded project (of mine) has ended up so far. I bet those guys have no idea what's going on and are taking guesses because no one understands the thing they've made.
mirsadm··on Small models also found the vulnerabilities that Mythos found
How is that going to find anything that interacts across files?
mirsadm··on System Card: Claude Mythos Preview [pdf]
The numbers only go up to 100% though.
mirsadm··on We rewrote our Rust WASM parser in TypeScript and it got faster
Not in my experience. They're pretty good at getting average performance which is often better than most programmers seem to be willing to aim for.
mirsadm··on An industrial piping contractor on Claude Code [video]
I've noticed even experienced engineers have started overestimating how long things would take to build without AI. Believe it or not we coded before AI and not everything took years all the time.
mirsadm··on If you thought code writing speed was your problem you have bigger problems
I don't know if I agree with that. It's a struggle to get a modern compiler to vectorize a basic loop.
mirsadm··on Coding after coders: The end of computer programming as we know it?
That is already the struggle. There is too much stuff already.
mirsadm··on The changing goalposts of AGI and timelines
Why? Give it a couple weeks and everybody will forget about this. They'll be earning more money than previously. Job well done.
mirsadm··on LLMs work best when the user defines their acceptance criteria first
100% dependent on the person driving it
Page 1 of 18Next →