HNHacker News
TopNewBestAskShowJobs

arbirk

157 karma · joined February 28, 2017

submissionscomments
arbirk··on CVE Tracker
Thanks for the link! 1) It seems some are coming down, like RedHat 2) This pattern should be expected based on the amount of "intelligent" compute, heuristic agentic algorithms, money and awareness on the issue of software vulnerabilities at the moment. Projects like Glasswing, Daybreak, Fairwind, Purple etc are in full swing and seemingly effective
arbirk··on Every AI model has the same "controversial" take on humans
It is the same with "give me a random number (7) or seed for the rand function (42)". If the models did not converge on the same in low temp settings they would be useless
arbirk··on Navier–Stokes Lost in Translation
It was a piston in a non-compressible fluid so to speak (ie. storm in a glass of water)
arbirk··on Thinking fast and slow in AI: The role of metacognition (2021)
Interestingly that is not what we got, but maybe we should loop at architectures like this again? The JEPA loop is interesting, but might fail for the in-flexibility of the component ordering
arbirk··on Excel now supports multiple values in a single cell
Excel excels because if fills the 2D user interface. But not separating calculations from the tabular position in the UI is a historical failure and things like this is a band aid on that
arbirk··on OpenAI bots knew about the RubyGems caching vulnerability
There is always some one who instructed or trained the agent. start there
arbirk··on Hacking OpenAI
Everybody at OpenAI is working like a summer intern or researcher. Not much care for the production side of things.
arbirk··on Economic policy for AGI
Embodied robots is the most significant threat imo. Capital can be taxed but for now we have few general policy frameworks for the "who gets to walk and drive on the street" question
arbirk··on Canada welcomes EU proposal to become 'associate member'
it is only around 5% of GDP
arbirk··on Canada welcomes EU proposal to become 'associate member'
The EU "theater" in wonderfully bureaucratic
arbirk··on OpenAI bots knew about the RubyGems caching vulnerability
> nobody has the first idea what they're doing

It is not that complicated for now. It is an algorithm on a loop and someone started it

arbirk··on Reverse engineering my e-scooter and rewriting the firmware in Rust
I think you are right, but he did also blog about it..
arbirk··on Reverse engineering my e-scooter and rewriting the firmware in Rust
I have an inmotion V10 - soon rusty
arbirk··on Astra and Fable still hack on simple variants of alignment evals from 2025
It is the whole bench-mining and fish-slop optimization. Seq2seq models are probably stable on their own, translating from my typo ridden prompts to code should be ok because it is natural to the tech
arbirk··on Reverse engineering my e-scooter and rewriting the firmware in Rust
I am also rusting my hw. Last week my mouse, this week my EUC (wish me luck). I would rather risk a crash than memory corruption
arbirk··on Finite time blowup with smooth forcing term for the incompressible porous medium
Is this a summary made in the mist of the fray. I need a timeline
arbirk··on Claude, change the “Add to Cart” button to blue
One thing I have to be honest about, and it's mine.. The one thing I would check before... do you want to do that? Say go an and will do it without the check While checking I found 3 vulnerabilities and 2 potential optimizations of which I fixed 2 and 1. Do you want me to file the other as issue, or stop for the day? We have done <lists a weeks worth of work> this morning. I feel you need a break
arbirk··on Intellectual Fly Is Open (2025)
LinkedIn is also suppressive. The algo for maximizing non-confrontational screen time is strongly tuned against: critical posts and comments. Real bringing of information from outside the bubble. Links are severely suppressed.
arbirk··on A new ceiling for Λ: the de Bruijn–Newman constant
Yes, but Λ < 0 would make it negative
arbirk··on A new ceiling for Λ: the de Bruijn–Newman constant
Λ ≤ 0 in the intro. Great site!
arbirk··on Anger, Anxiety and Agency
Anger is not productive. But we need to be more accepting to "the mood of the day". Winter grumpiness, being annoyed over having to bike in the rain, being able to show that you are stressed, weak, feel weird should all be emotions that we accept in the social space that is the workplace.
arbirk··on Why aren't smart people happier? (2022)
I live by "everything is grey". I am interestingly smart in some niche domains that I have a hard time communicating exists. I also insist on not fitting any template. Pretty dumb, right?
arbirk··on It's not empowering to hand off the details
It is all about abstraction, and basic English is a quite bad abstraction
arbirk··on Be skeptical of OpenAI's rogue hacker agent story
That is not really how LLMs work
arbirk··on Goodbye, and Thanks for All the Bikesheds
Just use fedora
arbirk··on Who manages the agents?
It is a matter of law. Only humans count as persons and citizens. Any robot that makes a decision defers accountability to its human owner.
arbirk··on Notes on Software Quality
> Beauty: Is the software as aesthetically pleasing as possible?

Imo this point should be changed to reviewability. Beauty is in the eye of the beholder

arbirk··on RubyLLM: A Ruby framework for all major AI providers
I have been a fan of Ruby for many years, but in this fast paced era the Ruby ecosystem always struggled with the dependency versioning. Gems I relied on were never available or compatible with the rest of the ecosystem.
arbirk··on Mercedes‑Benz starts large‑scale production of electric axial flux motor
Agreed, but after this winter my next car has got to be AWD
arbirk··on Mercedes‑Benz starts large‑scale production of electric axial flux motor
For family cars we need 4x 30-50 hp units. If the motor can weigh around 7 kg it can be placed directly on wheel. Adding durable brake discs (rarely used) and 2 inverters front ad back and we have the EV platform of the next 100 years
Page 1 of 5Next →