gaigalas

6 ms·

61.
by gaigalas 2mo ago
Recipe for delicious pizza: 1. Make extra greasy pizza. Nauseating-level of grease. 2. Put pizza through De-Greasinator 5000. Sounds efficient.
62.
by gaigalas 2mo ago
The fact that consumer-level DDR4 is getting more expensive makes absolutely no market sense to me.
63.
by gaigalas 2mo ago
If you need a full length blog post to explain it, it's not a good use of the word.
64.
by gaigalas 2mo ago
I was really excited about Altar I, but no shipping to Brazil, so I gave up on it. Nice to see the idea continues (consider worldwide shipping would be my only feedback on the idea, I guess). I now own a Logitech MX Keys S and I'm quit
65.
by gaigalas 3mo ago
Unironically, yes. It would be totally different.
66.
by gaigalas 3mo ago
Likely a 12 factor legacy issue. https://12factor.net/ This was really popular, and lots of people bought the factor III of only configuring via environment variables. The whole point of it was: keep config simple. Env vars
67.
by gaigalas 3mo ago
This is a well-known, established scholarly fact. https://en.wikipedia.org/wiki/Padapatha > The oral tradition of the Vedas (Śruti) consists of several pathas, "recitations" or ways of chanting the Vedic m
68.
by gaigalas 3mo ago
The recovery of a possible oral tradition regarding this is likely drawn from comparative mythology techniques, which is super interesting (I'm fascinated by it, personally), but in another category.
69.
by gaigalas 3mo ago
The Rigveda is the the earliest unbroken kept record. The Sumerian tablets are likely older, but they were forgotten and then re-deciphered. The transliterations still carry educated choices made by a scholar (there's only so much yo
70.
by gaigalas 3mo ago
Our earliest record is the Rigveda (a written down oral tradition, we don't know exactly how far back it goes). I suggest you try to read it at least once. It's almost incomprehensible.
71.
by gaigalas 3mo ago
A lot of Dunning-Kruger specialists here, apparently.
72.
by gaigalas 3mo ago
I think his explanation is bullshit. From my experience (PixiJS game), Opus can look at things and screenshot. It's slow but it works. The issue is that it doesn't try to make them look good. Instead, it tries to find one bug an
73.
by gaigalas 3mo ago
Funny how people are optimistic. The other giant growth leap mankind had happened just before the Bronze Age Collapse. Trade, automation, labor, same recipe. The outcome was a few centuries of dark ages though.
74.
by gaigalas 3mo ago
When the French used colorful obsolete uniforms in WWI, it was a skill issue. They had all the time in the world to modernize, but got cocky and made a terrible mistake. The US made their military supply chain too complex, and in several po
75.
by gaigalas 3mo ago
> Withholding technology is the only tool we have. The US has the strongest military in the world, tight control of a large part of the world economy, unprecedented cultural soft power, a permanent vetoing seat at the UN security counsel
76.
by gaigalas 3mo ago
I would accept something on the same magnitude in other areas. AI companies promised a lot of things in the beginning. "Cure cancer", "no one works anymore", "intelligence as cheap as electricity". Any of those
77.
by gaigalas 3mo ago
AI companies themselves pitched this. It's their word. Now no one talks about it (obviously), but I remember.
78.
by gaigalas 3mo ago
That addresses only half of what I said. That's why you assume I'm assigning fixed roles, you're not seeing the big picture. Can you go to a poor area and change it with a speech about ambition? Likely not that much. The issu
79.
by gaigalas 3mo ago
The people qualified to do great work probably don't need the advice. They most likely lack resources (like money or free time), not guidance. The people who would benefit from the advice (resourceful people full of free time that are
80.
by gaigalas 3mo ago
The only honest benchmark is "has it cured cancer yet?". It has not, nor moved close to doing so. People are doing great at marketing the diminishing returns as breakthroughs though.
81.
by gaigalas 3mo ago
One thing I mentioned before is that the representation of artifacts produced by AI being inspectable has notable value. I miss this in some existing agentic interfaces. I'm not typing into the text editor, I'm inspecting the work
82.
by gaigalas 3mo ago
That makes a lot of sense when reading, but in reality a lot of things have snapped back to plain text over and over again. I don't know the exact reason for that, but it's undeniable that it happened. Lots of people try completel
83.
by gaigalas 3mo ago
It does feel novel. However, I believe the GUI for AI agents that will win is a plain text editor with a file tree, a terminal pane for coding (or preview pane for other kinds of tasks) and a chat sidebar. Even if the code editor is not use
84.
by gaigalas 3mo ago
That second drawing looks more like a guy who has a cool coat full of trinkets inside than actual wings. Look how they're not actually feathers, but several distinct objects inside the coat. Like pendants, coins and rings (some of the
85.
by gaigalas 3mo ago
> Open source did more than lower the cost of software; it created a shared foundation of knowledge on which generations of American engineers and entrepreneurs built their institutional sovereignty Hey nvidia, what about making your ful
86.
by gaigalas 3mo ago
Confirmation bias. There's likely a wide portion of chats in which "keep going" derails to madness. We then stop saying "keep going" because we notice there is something wrong, and we start another chat. In the end,
87.
by gaigalas 3mo ago
(author here) index.sh is actually quite easy to read (in case someone wants to audit it before running), 800LOC pure shell + tuish deps, forkless design. https://github.com/alganet/alganet.github.io/blob/mast
88.
Bash for Browser Front End(alganet.github.io)
4 points by gaigalas 3mo ago|2 comments
89.
ML-pokedex – experiments trying to train some models directly on RDF(alganet.github.io)
2 points by gaigalas 3mo ago|0 comments
90.
by gaigalas 3mo ago
We're not seeing the big swing yet, which is when hardware becomes cheap enough so that hobbyists and volunteers can meaningfully contribute to a shared open substrate. We'll see smaller, more efficient models, better training and

More ›