Hacker News: adrianN

New comment by adrianN in "Qwen3.7-Max: The Agent Frontier"

adrianN — Wed, 20 May 2026 14:24:08 +0000

In an ideal world everybody runs open models on hardware they control.

New comment by adrianN in "Project Glasswing: what Mythos showed us"

adrianN — Tue, 19 May 2026 00:56:03 +0000

I’m not sure that having people read LLM output does that either.

New comment by adrianN in "Earth's Radio Bubble: Every signal we've ever sent into space"

adrianN — Tue, 19 May 2026 00:53:56 +0000

We don’t have a lot of examples of technological life forms that don’t need oxygen.

New comment by adrianN in "Project Glasswing: what Mythos showed us"

adrianN — Mon, 18 May 2026 16:43:34 +0000

Writing and later the printing press have already considerably stifled human expressiveness. Language used to be noch more fragmented and diverse before mass media (or the Bible in every household). In my grandmother’s time you would have difficulty understanding people from three villages down the road.

New comment by adrianN in "I don't think AI will make your processes go faster"

adrianN — Sun, 17 May 2026 16:47:15 +0000

I believe it is very hard to quantify „productivity“. I’m sure that for suitable definitions you can find gains from coding assistants. Personally I get more code written and more features implemented. Yet I’m very wary of coding assistants because I believe they deal a fatal blow to my ability to understand the system. All LLM generated code is (at best!) code that was written by an intern which I just helped with the design and reviewed (unless productivity expectations cut down my review time and I get LLM assistance for reviews too). My grasp on the inner working of that code is much more tenuous than had I written it myself. I will never become an expert by just reviewing code and prompting.

For a while this is not a problem: I can work with my current mental model. But every generated PR erodes my expertise a little bit. Eventually my mental model won’t fit anymore.

So how much of that model maintenance should I count into my productivity metric? Does that even matter or will the next model be able to reason well enough that my mental model doesn’t matter?

New comment by adrianN in "The main thing about P2P meth is that there's so much of it (2021)"

adrianN — Sat, 16 May 2026 06:47:15 +0000

I consider it the governments job to keep the population in a state where they can have a net benefit to society. Obviously this needs to be balanced with personal freedoms.

New comment by adrianN in "First public macOS kernel memory corruption exploit on Apple M5"

adrianN — Fri, 15 May 2026 06:44:24 +0000

I suppose that if you don’t believe that models will be good enough to work completely without senior engineer help, positioning yourself as a master prompter is a good move to improve your chances of not getting fired.

New comment by adrianN in "First public macOS kernel memory corruption exploit on Apple M5"

adrianN — Fri, 15 May 2026 02:47:06 +0000

The gamble is that you can cruise on the senior engineer’s diminishing understanding for a few years until models become good enough that you don’t need any humans in the loop and you can fire all those expensive seniors.

New comment by adrianN in "Deterministic Fully-Static Whole-Binary Translation Without Heuristics"

adrianN — Wed, 13 May 2026 15:54:11 +0000

That doesn’t matter. Once the code is generated it doesn’t change. The reviewed artifact in a safety critical codebase is the last abstraction layer before a fully certified compilation pipeline. So usually it’s not the UML but the generated code.

New comment by adrianN in "Deterministic Fully-Static Whole-Binary Translation Without Heuristics"

adrianN — Wed, 13 May 2026 10:39:13 +0000

There probably already is generated code in there, only it was generated from UML. I don’t think that LLM generated code will be treated differently from the point of view of the relevant regulations.

New comment by adrianN in "Deterministic Fully-Static Whole-Binary Translation Without Heuristics"

adrianN — Wed, 13 May 2026 08:26:54 +0000

Safety critical software is mostly a compliance dance that incidentally produces artifacts with lower defect rates than usual. LLMs can help with safety critical code as long as a human signs their name that they are responsible for its behavior.

New comment by adrianN in "I'm going back to writing code by hand"

adrianN — Mon, 11 May 2026 07:45:45 +0000

Does that matter that much in practice? I bet lots of costumers are okay with software that crashes 10x as much if it costs 10x less. There already is a ton of shitty software that still sells.

New comment by adrianN in "I'm going back to writing code by hand"

adrianN — Mon, 11 May 2026 07:43:25 +0000

To be fair, many of the proof attempts that mathematicians do also have major flaws. Most get caught before getting published.

New comment by adrianN in "France moves to break encrypted messaging"

adrianN — Sun, 10 May 2026 06:01:59 +0000

Politicians should probably not use Signal but something that is controlled by the government and for example doesn’t allow „accidentally“ deleting incriminating messages.

New comment by adrianN in "Poka-Yoke"

adrianN — Sat, 09 May 2026 07:14:33 +0000

Designing infrastructure and processes to reduce human error is a very important topic.

With the recent AI boom, many engineers are working on figuring out how to build infrastructure that reduces AI errors. We rediscover many techniques (eg writing good specs) that were once used for humans, but then abandoned.

New comment by adrianN in "The Old Guard: Confronting America's Gerontocratic Crisis"

adrianN — Thu, 07 May 2026 02:48:31 +0000

Social security is already a redistribution scheme, it already is more complicated.

New comment by adrianN in "The Old Guard: Confronting America's Gerontocratic Crisis"

adrianN — Thu, 07 May 2026 02:46:01 +0000

Demographics in most developed countries make catering to old people the strategy to win elections.

New comment by adrianN in "Colombia hosts talks on exiting fossil fuels as global energy crisis deepens"

adrianN — Thu, 07 May 2026 02:41:43 +0000

It is a bit sad that we have to celebrate bending the third derivative of the relevant metric in the right direction after like fifty years of talks.

New comment by adrianN in "Colombia hosts talks on exiting fossil fuels as global energy crisis deepens"

adrianN — Wed, 06 May 2026 18:20:47 +0000

I would argue that subsidized solar panels and batteries from China are the the most important factor. If renewables weren’t economically competitive we’d see approximately zero deployment.

New comment by adrianN in "Colombia hosts talks on exiting fossil fuels as global energy crisis deepens"

adrianN — Wed, 06 May 2026 18:08:56 +0000

We‘ve had talks about this topic decades before I was born, but progress is a bit underwhelming.