5,617 karma · joined August 30, 2017
Codeberg: https://codeberg.org/ashton314
GitHub: https://github.com/ashton314
Blog: https://lambdaland.org
As an aside, have you seen Typst? It’s got LaTeX-level typesetting quality but the markup syntax is a lot friendlier (close to Markdown) and the scripting language is a Real Language™ with sensible error messages and sub-second compilation times even for big documents.
But for less-excellent authors, where’s a good place besides Amazon to get ebooks?
Not possible for all problems. We cannot decide correctness (ie adherence to a specification) for all programs, but we can definitely recognize a good chunk of cases (both positive and negative) that are useful.
The Halting Problem itself is recognizable. The surprising result of Turing’s work was that we can’t decide it.
Loved the Project Hail Mary quote from one of the mission controllers. :)
This bright spot in world news has been good for my mental health and general motivation. Thank you NASA!
> If the issue is that people write bad floating-point expressions, a code-writing tutorial would be a better solution.
Yeah you are just criticizing this without even looking at it. Shame.
I use Zotero [1] to manage/read/annotate all my papers and it's got a built-in PDF inverter that works pretty well. I'll take Veil out for a spin some time and see if it works well in places where Zotero's algorithm fails.
[1]: https://zotero.org
Instead of "I understand the causal mechanism and can predict what happens if I change X," you get something more like "I have a sufficiently rich model that I can simulate what happens if I change X, with probabilistic confidence." The answers are distributions, not deterministic outputs. That's a different kind of knowing.
At the beginning this sounded like, "hard problems are complex, machine learning can help us manage complexity, therefore we will be able to solve hard problems with machine learning", which betrays a shallowness of understanding. I think what this essay argues here is a little deeper than that trite tech-bro hype meme.
But I disagree with this conclusion: I don't know that we can begin to build these models to begin with or that our new LLM/transformer-powered tools can help solve these problems. If simulation were the answer to everything, why will new ML tools make a significant difference in ways that existing simulation tools do not?
Stuff like AlphaFold is amazing—I'm not saying that better medical results won't come about from ML—but I feel like there's some substance missing and that even this level of excitement that the author expresses here needs more and better backing.
;; Speed up magit status by removing some things
(remove-hook 'magit-status-sections-hook 'magit-insert-tags-header)
(remove-hook 'magit-status-sections-hook 'magit-insert-status-headers)
(remove-hook 'magit-status-sections-hook 'magit-insert-unpushed-to-pushremote)
(remove-hook 'magit-status-sections-hook 'magit-insert-unpulled-from-pushremote)
(remove-hook 'magit-status-sections-hook 'magit-insert-unpulled-from-upstream)
(remove-hook 'magit-status-sections-hook 'magit-insert-unpushed-to-upstream-or-recent)In undergrad my iPad was far and away my favorite note-taking device. Digital pen-and-"paper" beats laptop for 99% of note taking.
[1]: https://veloren.net/
Maybe I should have been writing everything to a file. ¯\_(ツ)_/¯ Anyway, I didn’t think of it at the time and Ghostty saved me.
I hear you and you're right that Codeberg has some struggles. If anyone needs to host critical infra, you're better off self-hosting a Forgejo instance. For personal stuff? Codeberg is more than good enough.
Codeberg does suffer from the occasional DDOS attack—it doesn't have the resources that GH has to mitigate these sorts of things. Also, if you're across the pond, then latency can be a bit of an issue. That said, the pages are lighter weight, and on stable but low-bandwith connections, Codeberg loads really quickly and all the regular operations are supper zippy.
That's not lot, mate. Maybe more than the average American unfortunately, but I consider a year where I get through 2–3 books a slow one. And that's just reading I do outside of my job. What matters, of course, is both the quality and quantity of what you read. The short of it is, your attempt at building your ethos has fallen pretty flat here.
> AI is going to disrupt the whole academia [sic]
Yes, it has and it continues to. I'm not arguing against that. Joe Liemandt said that "all educational content is obsolete", which presumably includes not only textbooks like SICP or Sipser's Introduction to the Theory of Computation (just listing some CS textbooks because I'm in CS) but also great works of literature and philosophy that are important texts like The Odyssey or A Tale of Two Cities. If he meant to exclude such texts from the umbrella of "all educational content", well, then that's telling too. :)
> …it is infinitely better than a book or a teacher.
Maybe for someone who struggles with literacy or who hasn't had the pleasure of a good teacher. If you really believe this I'd like to see you try to substantiate your claim.
> The student could move at his/her own pace…
The article is about a grade school kids who, most of the time, need a little pushing to reach their full potential.
> …ask questions if stuck which no book or teacher could deliver.
… you're saying that LLMs are better than teachers because you can ask LLMs questions and not teachers?! Also, asking questions isn't the only component of learning. A good teacher will know when to not answer a question (or ask one!) and let the student stew and think about it.
I'm not saying AI can help with education. It can—it helps me!—but no hallucinating stochastic machine will have the human insight that a good teacher has. It's not a replacement.
I wonder if this fellow has ever read a serious book. I'm skeptical.
This! This is my feeling exactly. I wrote about encountering work slop last year: https://lambdaland.org/posts/2025-08-04_artifical_inanity/