531 karma · joined June 13, 2025
Gary Marcus’ takes have aged really poorly over time ironically he is the “AI expert” who has to constantly move the goal posts.
I.e. the AI 2027 guys were memed for being AI lunatics on a lot here and they have been pretty on the money in terms of pace of progress accelerating/ gov moving towards nationalization/ coding agents
> RSI within a top-down, concentrated power structure creates an unstable equilibrium and will instantly devolve into an epic power struggle. If you control the company/computers/military with the one most powerful thing, you control everything else
With a more stable government you can definitely imagine a situation where someone like Anthropic would work closely with the government in a Manhattan style project to solve alignment as soon as RSI is close. This admin definitely has a more "might makes right" view of the world and I'm not sure they would go that direction.
Not like I have any inside info.
I was mainly responding to this I know what you mean.
The reason I bring up OpenAI is because a lot of the discourse has boiled down to "both sides are the same" which I don't think is fair.
Here is Dario writing about AI safety in 2016 [1]. Dario and others in the Anthropic circle have long been associated [2] with the effective altruism movement, which whatever you think about them, they are very concerned about AI existential risks. Ronan Farrow & the New Yorker did a mega deep dive on Sam Altmans history of being dishonest and manipulative, and it credibly reports that Anthropic cofounders left due to not trusting leadership there with safety.
On the other hand we have decades long evidence of Altman and co. being dishonest and employing people like Chris Lehane [3]. I have no affiliation with Anthropic beyond being a user of their products to be clear but it feels like HN is going to end up on the wrong side of this one.
[1]. https://arxiv.org/abs/1606.06565 [2]. https://forum.effectivealtruism.org/posts/53Gc35vDLK2u5nBxP/... [3]. https://en.wikipedia.org/wiki/Chris_Lehane
I.e. OpenAI just went full evil corpo mode and went all in on the Leading the Future PAC [1] to try and prevent any kind of regulation.
I feel like there is a reasonable path where they might agree with OP that the government has "mostly gone insane" but also think that US getting its act together and leading the way on sane regulation will be key to getting to a good outcome with AI.
We start allowing supply again and the less desireable places on the commuter rail i.e: Brockton/ bridgewater/ Randolph get truly affordable over time, “top” destinations like Weston housing price growth stalls out at least.
I’ve never seen an area more adverse to economics 101 than Boston.
In the greater boston area NIMBY's are incredibly effective and will pull out every stop. I.e. instead of designating normal areas as multifamily, they will designate a portion of a golf course as multi family so you'd have to buy out the whole golf course [1].
It really is sad MA is 50/50 in housing produciton per capita [2]. Despite being severely under built and very desirable place to live, and then everyone pretends to be progressive. Michelle Wu in Boston uses "affordable housing" to bring all residential construction to a halt.
[1]. https://www.cbsnews.com/boston/news/marblehead-david-modica-...
No one really knows what the endgame of software security looks like.
So some people should try the port to rust angle, some should focus on hardening the C, some should explore more exotic options like formally provable languages etc
Basically I use these "kits" to prove that the behavior is working as expected with mocked data/ interfaces and then only after these kits pass I'll run the real test suite files as confirmation. So these let you iterate a lot faster than the official test suite because it is very slow.
These are bootstrapped from the real tests.
The other commenter was being a bit dismissive but this is the kind of thing I'm taking away as a real useful pattern to do verification of behavior at scale.
The project was inspired by that. Also unlike most other projects, nginx is directly exposed to the internet often times which makes it more vulnerable than i.e. Redis/ Valkey or something that would be running within a companies network generally.
"C Bad" is a bit reductionist... but I think there is some truth to the take " Until you have the evidence, don’t bother with hypothetical notions that someone can write 10 million lines of C without ubiquitious memory-unsafety vulnerabilities – it’s just Flat Earth Theory for software engineers" [1]
NSA and other government orgs are also pushing people to stop using C [2] for important software.
[1]. https://alexgaynor.net/2020/may/27/science-on-memory-unsafet... [2]. https://linuxsecurity.com/news/government/nsa-s-plea-stop-us...
But in terms of learning I'm learning relatively little about how to type Rust into an editor but a lot about how to set up agentic loops that can autonomously get tests to pass and improve performance.
For example if you just tell a frontier model (gpt5.5 or Claude Code 4.8) to make some portion of the tests pass they will take forever and just bang their heads against it. I developed a framework to mimic a lot of these tests in nginx... but in minimum non blocking ways so you can run many in parallel with short feedback loops.
Similar for performance - how to make tons of performance benchmark and expose maximum telemetry for agents to go and analyze the hotpaths etc.
And yes it seems like there has been many attempts to get a solid Rust Lua over the years and most never reached parity so hoping some people can find use case for it! This one is at full parity in terms of behavior and performance is getting to within striking distance.
[1]. https://github.com/ianm199/lua-rs/tree/main Lua
[2]. https://github.com/ianm199/valdr Valkey/ Redis
[3]. https://github.com/ianm199/nginx-rs-port nginx
Happy to answer any questions on the approach! When I started a few weeks ago the harnesses on their own were not good enough to get very far without a "meta harness" of sorts but that is changing largely with Claude Workloads and Mythos. A lot of the work is developing some custom tooling to move these along faster.
The other reason is that because mlua is just a wrapper around the C code, it has unsafe you can't really get around. So for example Lua is used in Redis, which has this critical CVE https://github.com/redis/redis/security/advisories/GHSA-4789... that a memory safe version of Lua wouldn't have to deal with.
Mlua is still fine or even better for many other cases though!
A small portion of this effort is having a high quality Lua in Rust repo. I’m using mythos to fix some of the performance issues with my Lua interpreter that gpt 5.5/ opus 4.8 had stone walled on.
Not sure if Mythos will be able to crack this but it has been running for a couple hours now with some promising results.
Performance charts linked here if your curious https://github.com/ianm199/lua-rs
> Any software producing stochastic or probabilistic output and cannot produce verifiable, repeatable, and predictable data cannot fundamentally replace something that requires a high level of proof and validation
Human output is also stochastic and probabilistic - our brains are not deterministic software. There is no fundamental reason we couldn't replace the human role in research with some form of AI (LLM or otherwise) if AI keeps improving.
That doesn't mean were going to go right to Dyson spheres so that every possible molecule is going towards scaling.
Isn't it simpler to believe that people like Terrance Tao who have been in good standing with the academic community for a long time are telling the truth? Like you may very well disagree with their predictions based on your lived experience or theories... But it doesn't take some crazy leap of faith to follow their logic of "AI could get better at fundamental/ applied STEM research than humans -> we can scale this up exponentially -> scaled up human level or super human level research will lead to major scientific breakthroughs".
Like sure CEO's like Amodei/ Altman/ and Musk may have an incentive to hype things up but not every opinion by every smart person has to be part of some grand conspiracy.