HNHacker News
TopNewBestAskShowJobs

kevmo314

3,232 karma · joined June 27, 2014

haha, computer go brrr

kevmo314@gmail.com

blog.kevmo314.com

[ my public key: https://keybase.io/kevmo314; my proof: https://keybase.io/kevmo314/sigs/jKuWflZVNiTzhIe0ttNcauvsm_uNCKaCk5U7Ni6SuFk ]

submissionscomments
kevmo314··on Show HN: Agent.email – sign up via curl, claim with a human OTP
Maybe if we destroy the internet faster we’ll move on to the next thing sooner
kevmo314··on AI coders are carrying half-open laptops through airports, offices, ice rinks
You can keep my computer's lid open for the low price of $25/mo
kevmo314··on The US is winning the AI race where it matters most: commercialization
There's a joke in China that Trump is the best president that China has ever had.
kevmo314··on MacBook Neo Deep Dive: Benchmarks, Wafer Economics, and the 8GB Gamble
Same, the Apple silicon chips have been huge.

I bought a 2019 Intel MBP and that was by far the worst laptop I've ever had. After just a year of use it was constantly overheating and running out of memory and disk space, barely able to open a terminal. It was so bad that I hesitated to buy the Apple silicon versions, but the good reviews convinced me and it has been going strong ever since.

kevmo314··on If AI writes your code, why use Python?
Should've posted to moltbook
kevmo314··on Cloudflare to cut about 20% workforce
That makes no sense. If that were really true then they could’ve saved a bundle and did that same thing years ago before AI.
kevmo314··on From Supabase to Clerk to Better Auth
Yes many people do say the same about AWS RDS.
kevmo314··on Does Employment Slow Cognitive Decline? Evidence from Labor Market Shocks
People don’t post about their daily 9-5 on instagram either.
kevmo314··on Show HN: VidStudio, a browser based video editor that doesn't upload your files
Wild that apps used to be completely local, no accounts, no uploads, and we're back to that as a value prop.
kevmo314··on 4-bit floating point FP4
Multiplication at this resolution is already implemented via lookup tables.
kevmo314··on Rust Threads on the GPU
Are you writing this from the future? The latest gen nvidia gpus sit at around 2-2.5 GHz and the latest gen amd cpus sit 4-5 GHz.

That matches my personal experience too, writing naive cuda code that doesn’t take advantage of parallelism is roughly half the speed of running it on cpu.

kevmo314··on The secrets of the Shinkansen
Ahh interesting. I wonder why they (non-Apple) did that?
kevmo314··on Rust Threads on the GPU
> a block of Rust threads that are properly programmed to take advantage of the vector processing by avoiding divergence

Sure, if you have that then of course it would be fast. But that’s not what this library is proposing.

kevmo314··on The secrets of the Shinkansen
Is this an Android thing? My US iOS works fine with digital Suica.
kevmo314··on Rust Threads on the GPU
Isn't this turning a GPU into a slower CPU? It's not like CPUs are slow, in fact they're quite a bit faster than any single GPU thread. If code is written in a GPU unaware way it's not going to take advantage of the reasons for being on the GPU in the first place.
kevmo314··on USB for Software Developers: An introduction to writing userspace USB drivers
If you are interested in doing this in golang I wrote a library to avoid needing cgo: https://github.com/kevmo314/go-usb

I use this to access UVC devices correspondingly without cgo: https://github.com/kevmo314/go-uvc

kevmo314··on Salomi, a research repo on extreme low-bit transformer quantization
Was this shoveled out with Claude/Codex to try to ride off the Bonsai release?
kevmo314··on Safeguarding cryptocurrency by disclosing quantum vulnerabilities responsibly
If someone else finds a cryptocurrency vulnerability, they too will reallocate as much of your allocation as they can and cash it out.
kevmo314··on CERN uses tiny AI models burned into silicon for real-time LHC data filtering
This might be some journalistic confusion. If you go to the CERN documentation at https://twiki.cern.ch/twiki/bin/view/CMSPublic/AXOL1TL2025 it states

> The AXOL1TL V5 architecture comprises a VICReg-trained feature extractor stacked on top of a VAE.

kevmo314··on 90% of Claude-linked output going to GitHub repos w <2 stars
Both of those statements are true.
kevmo314··on 90% of Claude-linked output going to GitHub repos w <2 stars
What is git if not a database for source code?
kevmo314··on Show HN: AI Roundtable – Let 200 models debate your question
It is funny that the AI's counterarguments amount to "you're hallucinating"
kevmo314··on Aliens.gov ~ domain registered 17MAR2026
Someone with the ability to register .gov domains is trying to make a sneaky buck.
kevmo314··on AIs can generate near-verbatim copies of novels from training data
Not a lawyer, but as I understand copyright is bound to distribution so if the person's perfect memorization of a book results in them reproducing it verbatim then probably yes.
kevmo314··on I found a useful Git one liner buried in leaked CIA developer docs
I don’t even trust myself to not mess up my git repo
kevmo314··on Show HN: Sameshi – a ~1200 Elo chess engine that fits within 2KB
I'm not suggesting that it's only brute force tree search, just that it's not very complicated to develop a theoretically perfect chess engine in direct response to the parent

> It's wild to think that 4096 bytes are sufficient to play chess on a level beyond anything humans ever achieved.

kevmo314··on Show HN: Sameshi – a ~1200 Elo chess engine that fits within 2KB
The core search algorithm is very simple though. 4KB engines may not run that fast if they do exhaustive search, but they’ll be quite accurate.

According to TCEC the time control is 30 mins + 3 sec, that’s a lot of compute!

kevmo314··on Another GitHub outage in the same day
I wonder if GitHub is feeling the crush of fully automated development workflows? Must be a crazy number of commits now to personal repos that will never convert to paid orgs.
kevmo314··on Eight more months of agents
Really? I hardly think it's neglected. The Claude Code harness is the only reason I come back to it. I've tried Claude via OpenCode or others and it doesn't work as well for me. If anything, I would even argue that prior to 4.6, the main reason Opus 4.5 felt like it improved over months was the harness.
kevmo314··on Company as Code
In the future, laying off half the company will be just one terraform apply away.
← PreviousPage 2 of 22Next →