Rewatched The Matrix the other night. The on-board food nozzles seem like an apt visual metaphor for ai slop.
13,077 karma · joined March 24, 2008
Location: San Francisco
Remote: Yesd
Willing to relocate: No
Technologies: TypeScript/JS, React, node.js, Redis, Postgres, pgvector, LLM agents, Cloud (AWS/GCP/Azure), ci/cd pipelines, other tech
Résumé/CV: on request
Email: charles@geuis.com
A few years later, it wasn't exactly programming but I was able to learn a bit about assembly on Macs in the 90s with ResEdit.
I almost get the sense you aren't a US citizen.
I still get value. But there's too much noise I think. I also think I'm older now and I'm more adverse to "done it, seen it"
The odd "made an app to do basic thing xyz" doesn't get my attention anymore.
Also, posts related to science get a hard scrutiny. Most posts are just cross references to poorly moderated subreddits. That, or they're just posts directly from archx with no external paper references. I'm not a scientist, but I know enough to know when to ignore bullshit.
Most posts to HN are still mostly bullshit. But I know enough to occasionally see something interesting. That's may be 4 times a year at this point.
...
I also want to add more context. I've been in the Anathem ideology for about 20 years now. It has shaped my world view. Call me fra geuis.
If you want to present what you think are new ideas, be prepared to back them up. When we get the Loreites up and running, I'll refer you to them.
Add ad tracking etc so what is this adding to the value of my life? You hope to get bought in a few years because you have a dataset (they) already have?
That also brings to mind that first exposure to this dataset affects the effectiveness of the rest of the dataset. If you're doing initial exposure, you'll definitely get the "aha" moment. But if all of the images in the dataset are of the same type, your brain quickly learns the pattern and the "aha" moment vanishes.
If they did their study on all of the images per test subject, the results after maybe the first 5 are basically useless for any definitive conclusions.
If anything, a solid counter argument can be made that Romance languages (descended from Latin) lack the flexibility of English and other Germanic languages.
Non primary English speakers frequently complain that English is more complicated than other languages. This is true. I'm a native speaker and only can read limited Spanish. Where I get hung up is the dependence on gender of objects. Similar experience with Japanese when I was studying that a few years ago.
I completely believe that primary language has a physical effect on the brain in terms of neural structure. It must have.
But since English is so adaptable, if there's a concept that is better expressed in another language we tend to adopt the words of other languages to express it.
However other languages seem to be less adaptable. For example, France has or had an official government ministry for decades to manage new foreign words entering the French language. To this day, there are newish specific French words for technologies coming from English speaking countries.
Another good example is some YouTube videos from India I've run across. (I turn on subtitles). But say the speaker is talking in Hindi. Many times more technical terms are English words or phrases that are freely interspersed with Hindi. They're borrowing the English words, with a bit of a Hindi dialect hitting the pronunciation.
Going back to Japanese, we see the same thing. I don't know if the JP gov has a language ministry.
But if you look at written Japanese text you definitely see that most numerology is written with western/English 0-9 characters mixed with katakana or hiragana. When you hear people speaking, and once your ear is oriented towards Japanese sounds, you can start to pick up on the adopted English words that are said with a native dialect emphasis.
As a side feature, that would be neat though.
Within like an hour or two pg emailed me asking me to stop. I didn't know it at the time, but HN was being run on a rusty potato and scraping the homepage every 5 or 10 seconds was causing significant load.
Location: San Francisco
Remote: Yes
Willing to relocate: No
Technologies: TypeScript/JS, React, node.js, Redis, Postgres, pgvector, LLM agents, Cloud (AWS/GCP/Azure), ci/cd ipelines
Résumé/CV: on request
Email: charles@geuis.com
Current gen AI is going to result in the excess datacenter equivalent of dark fiber from the 2000's. Lots of early buildout and super investment, followed by lack of customer demand and later cheaper access to physical compute.
The current neural network software architecture is pretty limited. Hundreds of billions of dollars of investor money has gone into scaling backprop networks and we've quickly hit the limits. There will be some advancements, but it's clear we're already at the flat part of the current s-curve.
There's probably some interesting new architectures already in the works either from postdocs or in tiny startups that will become the base of the next curve in the next 18 months. If so, one or more may be able to take advantage of the current overbuild in data centers.
However, compute has an expiration date like old milk. It won't physically expire but the potential economic potential decreases as tech increases. But if the timing is right, there is going to be a huge opportunity for the next early adopters.
So what's next?
My workflow is hard to describe.
Over the years I updated my base project files when new Blender versions require rebuilding my geometry node graph when old stuff gets outdated. That being mentioned, the updates are worth it. It's a hobby so time is free and a learning opportunity.
It's something I accept as cost of learning, which so far has resulted in better end results.
I was totally in the same mental space with Fusion. Here's what I did.
I don't like hard to understand UI.
I don't like a product whose company constantly changes the rules.
I don't like not being in specific control of mission important software I use.
I also tried open source alternatives to CAD. There are non that are anything approachable from a user perspective. Until the FreeCAD project gets some help from Blender, I'll stand by that.
I don't maybe have any specific modeling scenarios to recommend. Not a professional, just a hobbyist.
I tune my printer. Use the 3d toolbox plugin for Blender. Make sure your model is manifold. Get to learn how to spend dozens of hours editing 3rd part "printable" models into something that's actually printable. Ain't manifold, ain't printable. If you "borrow" models from games to print, you'll spend a lot of time making them printable. They aren't yours even at that point. Don't try to sell them or give them away for free. Not yours. Respect the artists.
Geometry nodes have SOOOOO many options. I'm not kidding, it's awesome the team keeps adding nodes that address old and new issues. But figure out a basic workflow wherever you're at. Only update if there is some required or otherwise very specific advantage
I probably spend more time for some projects in blender than I would in a proper cad environment, but it's a toolset I understand somewhat deeply now.
So far, Blender is one of the most successful open source projects I've ever worked with in the last 30 years. Even with major changes, every change has mostly been for the better. The introduction of geometry nodes completely changed my workflow a few years ago. So much room for activities after that.
Unlike any other modern software package, I have no regrets or doubt when I upgrade to the latest Blender release.
If you're on the fence about picking up Blender I highly recommend taking the deep dive.
It has never translated well to custom font settings from a user browser, but it turns out those are exceptionally rare vs people who just ctl/cmd +/- on the keyboard or pinch and zoom on the phone.
I still remember the star selector pattern to prevent IE6 from reading malformed css. So me old.
But having grown up in my early career during the transition from embedding styles into html to sites like css zen garden, I am extremely critical of modern "frameworks".
I still see so much bloat in poor performing pages and often when I dig deeper it's tailwind in addition to other inefficiencies adding repeated cruft that doesn't help with site rendering.
A proper 20 year old understanding of css class style inheritance isn't hard, and it negates 80-90% of the complexity of maintaining complex web apps (react components mainly).
CSS grids are cute and useful, but for a hell of a lot of web interfaces a simple float and clear:left still works wonders.
More importantly, this video is too short. It doesn't talk about why this paper can last a thousand years other than saying that it has.
Also I think the UK host gets short shrift. I really want to know his story. That on its own would be super interesting.
If you're up for it, shoot me an email at charles@geuis.com.
- Ignore AI voiced books
- Show me unread books in series that I have in my library
- Experimenting with better search. I have experience with building semantic search systems and have been highly disappointed with Audible's extremely sub-par search capabilities. I want results that are actually based on books, authors, and narrators that I have already purchased, read, or listened to.
- Get automatic notifications when new books from authors and narrators that I subscribe to become available.
There's at least a few more gripes I want to address, but these are the high priority ones that come to mind right now.
This was the venture funding "we're a startup era". And Google succeeded eventually.
But in that era making money didn't matter. It was just about grabbing market space. And oh boy did they succeed.
But all bills become due eventually. Stock holders start demanding continuing increasing profit and that eventually leads to the downfall of any good product.