Hacker News: johngossman

New comment by johngossman in "Marc Andreessen is wrong about introspection"

johngossman — Fri, 03 Apr 2026 16:19:06 +0000

In the late 1990s I went to a RealNetworks developer conference and Andreesen, then at Netscape, was a keynote speaker. I was curious and open to his insights, but his talk was so vapid (I remember he kept giggling) and arrogant that I eventually walked out. I remember he kept bragging about Netscape's next big project (something after Netscape 5 maybe?) and how it was going to wipe Microsoft out permanently. Only a few years later did I realize whatever it was never shipped, it turned out to be vaporware.

New comment by johngossman in "CERN uses ultra-compact AI models on FPGAs for real-time LHC data filtering"

johngossman — Sat, 28 Mar 2026 13:43:01 +0000

Not really. Think of the experiment as a very, very high speed camera. They can't store every frame, so they try to capture just the "interesting" ones. They also store some random ones that can be used later as controls or in case they realize they've missed something. That's the whole job of these various layers of algorithms: recognizing interesting frames. Sometimes a new experiment basically just changes the definition of "interesting"

New comment by johngossman in "Last gasps of the rent seeking class?"

johngossman — Fri, 27 Mar 2026 16:26:49 +0000

Like self-driving taxis where the business model is to stop paying drivers so we can pay more to big tech companies. Viva la revolution!

New comment by johngossman in "Mystery jump in oil trading ahead of Trump post draws scrutiny"

johngossman — Tue, 24 Mar 2026 17:46:02 +0000

It just feels that way sometimes

New comment by johngossman in "Mystery jump in oil trading ahead of Trump post draws scrutiny"

johngossman — Tue, 24 Mar 2026 16:38:50 +0000

"Almost every" is a very strong statement. But even granted that, the interregnum periods (civil wars and revolutions) tend to be so horrific that they are wise to avoid. In fact, people like Plato, Machiavelli, and Hobbes who lived through revolutions tended to come to the cynical conclusion that any system of government was better than a civil war. I don't agree with that conclusion, but I'd rather see the system reform itself than jump immediately to "tear up the constitution and start over"

New comment by johngossman in "Five years of running a systems reading group at Microsoft"

johngossman — Sun, 22 Mar 2026 20:41:41 +0000

In 35 years in the industry, reading and studying during work hours were always supported. Frankly, most places would let us play video games during work hours as long as we met our deadlines.

New comment by johngossman in "Leviathan (1651)"

johngossman — Wed, 18 Mar 2026 13:14:04 +0000

Machiavelli. Not just the Prince but his other works. He reads remarkably modern. There are many "Machiavelli Readers" that will provide a curated selection.

Cicero and Plato.

New comment by johngossman in "Corruption erodes social trust more in democracies than in autocracies"

johngossman — Mon, 16 Mar 2026 15:45:37 +0000

You might reconsider when your richer neighbor paid the politician to block you or build an asphalt plant next to your new condo. It's a slippery slope. Or how about when the fire department starts asking for a little something to keep your condo "safe"

New comment by johngossman in "The Bovadium Fragments: Together with The Origin of Bovadium"

johngossman — Fri, 13 Mar 2026 16:19:23 +0000

Tolkien and Lewis came by their luddism fairly, having both survived the horrors of trench warfare.

New comment by johngossman in "Verification debt: the hidden cost of AI-generated code"

johngossman — Sat, 07 Mar 2026 18:24:58 +0000

Yup. I've had Gemini create fake citations to papers. I've also had it hallucinate the contents of paywalled papers, so I know I can't trust anything it writes, though I am getting better at using it recursively to verify things.

New comment by johngossman in "Verification debt: the hidden cost of AI-generated code"

johngossman — Sat, 07 Mar 2026 17:54:20 +0000

This verification problem is general.

As an experiment, I had Claude Cowork write a history book. I chose as subject a biography of Paolo Sarpi, a Venetian thinker most active in the early 17th century. I chose the subject because I know something about him, but am far from expert, because many of the sources in Italian, in which I am a beginner, and because many of the sources are behind paywalls, which does not mean the AIs haven't been trained on them.

I prompted it to cite and footnote all sources, avoid plagiarism and AI-style writing. After 5 hours, it was finished (amusingly, it generated JavaScript and emitted a DOCX). And then I read the book. There was still a lingering jauntiness and breathlessness ("Paolo Sarpi was a pivotal figure in European history!") but various online checkers did not detect AI writing or plagiarism. I spot checked the footnotes and dates. But clearly this was a huge job, especially since I couldn't see behind the paywalls (if I worked for a Uni I probably could).

Finally, I used Gemini Deep Research to confirm the historical facts and that all the cited sources exist. Gemini thought it was all good.

But how do I know Gemini didn't hallucinate the same things Claude did?

Definitely an incredible research tool. If I were actually writing such a book, this would be a big start. But verification would still be a huge effort.

New comment by johngossman in "New evidence that Cantor plagiarized Dedekind?"

johngossman — Sat, 28 Feb 2026 18:33:19 +0000

I find they are good at identifying interesting topics and writing articles that don't deliver. They remind me of Omni magazine (which I subscribed to at one point). The articles aren't even wrong.

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 17:42:26 +0000

And reading a history book while you're in the place makes both the book and the place more interesting.

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 17:41:03 +0000

One more source to consider, if you find non-fiction dry (and if you do, I'd start by trying a different history book) is novels. Novelists like Rushdie, Heller, Solzhenitsyn, and a surprising number of sci-fi writers include a lot of history in their books. Clearly, these are not unbiased sources and can't be relied on as your only source, but often make for easy insightful introductions and/or immersive supplements. I'm not even talking about "historical fiction," though some of that is also good. For example, the book "The Killer Angels" about Gettysburg, is sometimes assigned reading in history classes.

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 17:31:15 +0000

I wish I could upvote this multiple times. However, I hope nobody takes away from this that all history is purely biased and therefore worthless. Reading a history book should be like having a conversation with a well-informed friend: you should expect them to get most things right, talk with others, and feel entitled to your own opinion. But you shouldn't ignore them because they might be wrong or stop talking to them because you differ with them, and definitely not assume you're smarter and more clear thinking than the expert.

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 16:58:26 +0000

Are you seriously suggesting that one can't learn Roman History without learning to read Latin? That it is better to be ignorant of history than read a few books?

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 16:04:09 +0000

While I agree with the list of recommended podcasts, you may be a reader. It is easier for me to make notes and backup and review when reading. A good history book is a story book as interesting as any novel, with the added benefit of being about real people and events. Fortunately, there are a lot of good popular history authors, but unfortunately there are a lot of bad ones too. I looked at Goodreads and this list is pretty good:

https://www.goodreads.com/shelf/show/world-history

New comment by johngossman in "Those who can, teach history"

johngossman — Thu, 26 Feb 2026 15:52:03 +0000

For someone not wishing to be an expert or read source material, like the OC, facility in languages is absolutely not necessary. And history absolutely is valuable and can be accessible to everyone.

New comment by johngossman in "Shatner is making an album with 35 metal icons"

johngossman — Tue, 24 Feb 2026 01:36:40 +0000

As I age, I look on these happy, productive seniors, people like Dick Van Dyke (100), David Attenborough (99), and Mel Brooks (99) and keep my fingers crossed.

New comment by johngossman in "Shatner is making an album with 35 metal icons"

johngossman — Tue, 24 Feb 2026 01:31:07 +0000

That's not a bug, that's a feature