geekraver

Counterpoint: there are plenty of companies making VR hardware but they all remain niche and struggle because they can’t subsidize their hardware by taking both losses and user data as part of the price. Meta are better at killing competition more than they are at innovating.

I am self-hosting with Gitea. Have action runners set up on my Windows PC and Mac (with a container for Linux). Works great.

AI is a tool. You can use it to make better software, or worse. In an ideal world we would use it to make better software; there’s enough bad software out there already.

Unfortunately, unfettered capitalism has few incentives to support that, and would rather we pushed out software faster with fewer people being paid less. Paying down tech debt has always been challenging.

So is the problem AI, or capitalism?

And when the heavy subsidies of cloud AI come to an end? Plus the Macbook can do other tasks as well; presumably you need a computer anyway so should deduct the value of a computer for non-AI tasks from the cost.

What about non-Finnish sauna heat? Or do I have to go to Finland?

Wow, totally ignores the one framework Microsoft developed that was actually a pleasure to use (WinForms). Avalon sucked IMO; it was just a continuation of the fever dream at the time of making everything XML (or XAML in that case). MFC sucked because it was C++. WinForms was that brief Window of “modern programming language with simple, elegant GUI framework”.

I know it’s not a popular opinion, and I am sure there were reasons Microsoft abandoned it, but that was a brief few years when I actually enjoyed building GUIs on Windows.

Because it’s a functional society that has a government that works for the citizenry and not a small cadre of wealthy sponsors?

ETTD

I do find it encouraging that there’s not a lot of pro-Trump sentiment on HN these days. Unlike, say, in October/November 2024…

Make sure to test you thyroid hormone levels

“Completely preventing X is impossible. As such, attempting to stop it is a foolish endeavor” has to be one of the dumbest arguments I’ve heard.

Substitute almost anything for X - “the robbing of banks”, “fatal car accidents”, etc.

Given the number of times Musk has been pissed or embarrassed by Grok saying things out of line with his extremist views, I wouldn’t be so quick to say it’s not intended. It would be easy enough to strip out of the returned system prompt.

And models eat apps over time. If you build an app that’s valuable, OpenAI takes note.

Same, mine is about to graduate with a CS masters from a great school. Couldn't get any internships, and is now incredibly negative about ever being able to find work, which doesn't help. We're pretty much looking at minimum wage jobs doing tech support for NGOs at this point (and the current wave of funding cuts from Federal government for those kind of orgs is certainly not going to help with that).

This is one of those places where having a selected “security image” makes sense. That popup should not just be easily spoofed text.

The pylance team has started exploring this, namely whether it makes sense to have an API for type checkers that is not the LSP, as language servers have a somewhat different goal in which type checking/inference is an enabling technology. This could allow multiple different language servers to be built on top of different type checkers (and the type checkers can run out-of-proc, so implementation languages can be different). https://github.com/microsoft/pylance-release/discussions/718...

Apart from Forth, Lisp and Lua, Tcl also deserves mention as a small embeddable language. Not sure of its state today; I think by v8 it had started to get more complex, but its original incarnation was very simple. I remember getting Ousterhout’s book and being so enamored with the elegant design that I felt compelled to just implement it myself from scratch, which took just two days.

“It can’t happen here”

Ph.D. programs being stopped: https://www.wesa.fm/health-science-tech/2025-02-21/universit...

Much better than some of the garbage solutions I have seen, including from sources that should know better, like The Algorithm Design Handbook. Some really absurd approaches out there, so bad I wrote a blog post about it in 2015: https://www.grahamwheeler.com/post/sudoku/

Friends don’t let friends use 4Chan - I mean X

https://www.audiolabs-erlangen.de/resources/MIR/FMP/C0/C0.ht... might have something

Is this going to be followed up with a “functional programming is the bestest” post?

Which they can avoid by resigning. So the worst penalty for almost any crime is they might have to step down,

So the president can’t kill their opponent over a golf game but can order the military to kill their opponent over a golf game. Got it.

In the worst case they might have to resign before being impeached but at that point they are permanently immune.

Germany 1933.

Along with K&R (and K and Plauger’s “Software Tools”), the Dragon book, Bentley’s Programming Pearls, and Holzmann’s Beyond Photography, AST’s books were the most formative in my life (I started coding in 1976 but was self-taught until early 80s when I got to college and read all these brilliant works). Long overdue recognition; so many people benefited from the lucidity of these minds.

This squares with my experience in the demo. It’s 98% there but that 2% is adding a lot of cognitive load and stress, so it’s a negative ROI.

Or when you have a #$@^$! notch

Not with standard LLM models. Generating garbage is easier; deciding that it’s garbage much more difficult. This is an inherent problem with “prompt optimizers” like DSPy.

This is cool. I have been doing this a bit more manually, by using a Chrome plugin that does YT summaries and shows transcriptions using Claude. I don’t like those summaries so I paste the transcriptions into ChatGPT (GPT4) with a prompt “Provide detailed study notes of the following video transcript”. That gives me a very similar format to yours. Will have to do some side-by-side comparisons.