aappleby

XTREEEEM

This was lovely.

Surely "how to draw a SVG pelican on a bike" has made it into the training data by now ...

tiny coin cell, would run for about six hours

You might enjoy this dime-sized earring I designed ~13 years ago - https://www.youtube.com/watch?v=XyBhdRp_5rY

this would be more convincing if it wasn't ai written.

The smallest amount of simple code that solves the problem wins. Everything else is irrelevant.

I feel like there's gotta be a catch in there somewhere, like "fails catastrophically when computing logarithms" or something. skeptical

Are you sure you're asking the right questions?

Glad I bought a JetKVM, it's been great (other than the HDMI adapter, lol)

No benchmarks?

Yes, but it doesn't need any funny parsing trick to handle them. Just parse the whole statement as a list of expressions joined by operators, and then you can convert the flat list into a precedence-respecting tree with a few lines of code and an operator-to-precedence table.

Why do I get the overwhelming feeling that the author is not a technical person and they had a LLM write this based on some handwavey ideas? There's virtually no _there_ there.

"...demonstrates its capabilities through worked examples" - The hell it does, your "examples" are three lines long. If you're going to compare it with LLMs, then have it do something LLM-ish. Or hell, the MNIST number recognition task would be better than the "hey look i modeled a flip-flop in my funny language" example.

Am I being harsh? Yes, I am. The author is claiming that they have a system that can automatically generate code for "quantum" and "spintronic" computers, yet offers zero proof of that.

You're probably already using a RISC-V computer, it's just embedded as a supervisor in some other gadget (or vehicle) you own.

Why "mediocre"? I've written production assembly language for a half-dozen different processor architectures and RISC-V is my favorite by far.

The article title is super misleading - this is about measurements of inflammatory markers in vitro and explicitly does not generalize to food intake.

No, the article title is misleading. This is in-vitro research only.

I'm late to the thread, but I was able to solve this on a microcontroller ~13 years ago.

https://youtu.be/yItm-9xl0as?si=9I4DLA3qETnQ1N2G

I get the overwhelming feeling that the author was convinced by an AI that he had a good idea without ever contacting an actual graphics programmer.

"Retro Game Engine owns the full frame lifecycle." - This is completely meaningless. Your engine controls whatever data is in the buffer that's sent to scanout, but the operating system and the GPU drivers and the scanout hardware in the GPU and the input processing and row/column drivers in the display control everything else.

The only actual screenshots this guy has are some "multiply the image with the subpixel mask" demos that.... don't look anything like a real CRT, and certainly nowhere near modern CRT shaders like CRT Royale.

The rest of the posts in the substack page are similarly devoid of actual content, but very heavy on the AI woo-woo this-is-important-and-deep stylings that I've come to find nauseating.

Hey Author, if you can see this - You're clearly a smart guy, but you need a basic grounding in 3D rendering if you're gonna do weird stuff - more than an AI can give you. In particular, the phrase "Light is linear" will be useful to you. Good luck.

Buried in this paper is a mention of "Applicative State Transition Systems", a programming paradigm that never caught on.

You wrote and shipped this in three days, eh?

hello chatgpt

This is... not an example of good optimization.

Focusing on micro-"optimizations" like this one do absolutely nothing for performance (how many times are you actually calling Instance() per frame?) and skips over the absolutely-mandatory PROFILE BEFORE YOU OPTIMIZE rule.

If a coworker asked me to review this CL, my comment would be "Why are you wasting both my time and yours?"

OpenTitan _is_ a microcontroller, just one with a _lot_ of security hardware (and security proofs).

It's intended to be integrated into a larger SoC and used for things like secure boot, though you could certainly fab it with its own RAM and GPIO and use it standalone.

Fiiiiinally! Yay!

Worked with the OT team at Google years ago and am glad to see this stuff finally taped out.

Ah, I thought you meant you were running it locally.

What are you running that model on?

I don't quite understand why the author is doing special handling for PSG versus PCM audio.

My GameBoy emulator generates one "audio sample" per clock tick (which is ~1 mhz, so massive 'oversampling'), decimates that signal down to like 100 ksample/sec, then uses a low-pass biquad filter or two to go down to 16 bit / 48 khz and remove beyond-Nyquist frequencies. Doesn't have any of the "muffling" properties this guy is seeing, aside from those literally caused by the low-pass.

This article is taking some liberties with the word "prodigy" that I disagree with.

If the way you nurture a talented student is via "intense drilling", I would argue that the student is not a prodigy in the traditional sense, but a talented and determined student who may or may not be dealing with parental pressure.

The actual prodigies I've known absorb information and gain skills without significant effort - I knew someone who enrolled in a calculus class, skimmed through the book in a week or so, and then would only show up to class for tests (which they would ace).

So the article conclusion doesn't surprise me - inflict relentless training on a young talented person and yeah maybe they won't want to do that as an adult.

But as far as actual "prodigies"? There is no burn-out because there is no (or minimal) effort. The choice of whether to stick with an area of interest through adulthood is more of a personal preference than anything ingrained.

What if the tool required an "un-safeword" to do destructive things?

"Do you really want to 'rm -rf /'? Type 'fiberglass' to proceed."