HNHacker News
TopNewBestAskShowJobs

mirekrusin

5,755 karma · joined September 4, 2014

https://mirekrusin.com
submissionscomments
mirekrusin··on Buz – A fork of Bun using modern Zig, with sub-1s incremental builds
You forgot to include link to your tool.
mirekrusin··on Startup founders urge U.S. government not to shut off Chinese open weight AI
They should codename next open model "flux capacitor".

Capability to travel in time is real issue as well as ability of open models to easily attack fable and sol - models which are completely defenseless.

mirekrusin··on Why do AI company logos look like buttholes? (2025)
OpenAI is more... open. Maybe that's what they wanted to communicate.
mirekrusin··on Kimi K3: Open Frontier Intelligence
They're not training on your data, they're training on "please anonymise this conversation" data.
mirekrusin··on Grok Build is open source
Unsubsidised token prices that people would actually pay for is a fantasy.
mirekrusin··on I tricked Claude into leaking your deepest, darkest secrets
Not paying anything feels off – it should be more evaluated against making it public information at the time of discovery until ie. public patch release, it doesn't feel right that the response is "trust us bro, we knew about it, bye", wouldn't hurt to drop some usage credits at least.
mirekrusin··on I tricked Claude into leaking your deepest, darkest secrets
Sound legit to me as long as it's prompted to use hardcoded "dario amodei".
mirekrusin··on Codex scraped the ICM website and discovered 2026 Fields Medal winner list
Isn't leaking browser extension used by one of people on the team (doesn't need to be developer, could be qa or anybody with whom the access was shared) more plausible?
mirekrusin··on How the terrorist group Boko Haram uses frontier AI
And now you’re ready to level up with motorbike jump.
mirekrusin··on SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence
It's simple: close weights = not welcome.
mirekrusin··on SWE-1.7 Reach Near GPT 5.5 and Opus Intelligence
Open weight models should have GPL-like license where it says if you train model on it, it needs to be open weight as well.
mirekrusin··on Study: ultra-black coating could reduce satellite light pollution
Give it few years and they'll get LED colors and we'll see them in "I'm lovin it" formation.
mirekrusin··on sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25)
it's not overpriced but high-priced looking from other countries and mostly not a markup, normal considering median wages.
mirekrusin··on sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25)
In others it's pizza night for family or half a bill for sushi dinner, so what?
mirekrusin··on ACL 1.0: A source-available commercial license for the AI era
What's the difference between this and Public Domain/CC0 from PoV of labs and vibers?
mirekrusin··on The gap between open weights LLMs and closed source LLMs
Closed weight models are a result of philanthropy by some private investors.
mirekrusin··on U.S. government will decide who gets to use GPT-5.6
"for the benefit of all humanity *"

* with big pockets

mirekrusin··on There is minimal downside to switching to open models
Banning models in US just strengthens competing states, ie. China.
mirekrusin··on NSA director: 'Mythos "broke into almost all of our classified systems in hours"
Yes, exactly, quite shocking, if something like this is true, as NSA (!!) director you keep it quiet, right?
mirekrusin··on NSA director: 'Mythos "broke into almost all of our classified systems in hours"
Mythos and other models are not brute-forcing passwords (and with this analogy passwords, ie. systems are the same).

We're not talking about dogs, but LLM systems.

Mythos is not exploring entire solution space either.

Usually looping is solved by repetition/frequency/presence/n-gram penalties/DRY/min-p sampling, not temperature but we're not talking about small models that have those classes of issues here.

mirekrusin··on NSA director: 'Mythos "broke into almost all of our classified systems in hours"
If mythos can break into almost all of their classified systems in hours then other models including opus, gpt, gemini and large open weight models can do so as well, maybe you'll have to double hours or it may become days, but they also will, there is no "maybe" in here.

State sponsored, non-public penetration fine tunes (of possibly public ones) likely can do it even faster.

Unsupervised penetration RL loop is ideal setup similar to optimization one – it's relatively easy to gain function on it.

mirekrusin··on Excessive nil pointer checks in Go
This suggestion fails for values that can be null, need to be mutable or need references from multiple places etc – it's not "just performance penalty".

Go has a problem, "just remember to always do X, never Y" patterns can't be guaranteed across all libraries you use, can't be enforced, can be violated for good reasons, other patterns and as a mistake etc etc.

Shame because otherwise it's a great language, but some mistakes are just no-go.

So close indeed.

They need Go 2 with *T and ?*T - that would be nice language to use.

mirekrusin··on Bun has an open PR adding shared-memory threads to JavaScriptCore
Did they rewrote WebKit in Rust or not yet?
mirekrusin··on LLMs Are Complicated Now
That reply was AI written.
mirekrusin··on GLM-5.2 is the new leading open weights model on Artificial Analysis
Did I get it wrong or the first link has dataset with 30 entries only?
mirekrusin··on GLM-5.2 is the new leading open weights model on Artificial Analysis
Wikipedia is a lot of data as well but we manage to do it, no?
mirekrusin··on GLM-5.2 is the new leading open weights model on Artificial Analysis
I don’t understand why there isn’t public dataset for reasoning that can be improved by humans/llms like Wikipedia (ie with auto judging contributions etc).
mirekrusin··on Show HN: High-Res Neural Cellular Automata
So the goal is to evaporate it with minimum number of shots?
mirekrusin··on Humanity isn't ready for the coming intelligence explosion
It's huge force multiplier for people who have hands on, detailed technical knowledge as well.

3 * N < 42 * N

42 - 3 < N * (42 - 3)

It helps to know layer you're working on.

People seem to make mistake of thinking how good LLMs are around tasks that they are familiar with and extrapolating it to whole population.

It's good mental exercise to think about how little you can do compared to expert on tasks you never thought of working.

Ie. if you're programmer or know something about finance, don't think how much it enables you to do better coding or investing, think instead how much it doesn't enable you to work on something you don't know like maybe molecular biology or visual special effects – it's all there but it's much better multiplier for people who do know their shit.

Knowing layer you're working on helps a lot, it gets multiplied.

Knowing programming is becoming more fundamental skill than ever before as it lies at the foundation of almost everything else.

mirekrusin··on SpaceX to buy Cursor for $60B
For comparison couple of months ago I acquired zed, pi, opencode etc. for $0.
← PreviousPage 5 of 34Next →