You know how certain people prepare and pack their basement for a major catastrophe?
Well, that's what I've been doing by making sure that when AI takes control of everyone's code, I can still program on my computer.
Dario Amodei on open source models:
“If the path continues, I think we could get to a very dangerous place.”
“When you control a model and you're deploying it, you have the ability to monitor its usage.”
“You can revoke a user's access. You can change what the model is willing to do.”
“When a model is released in an uncontrolled manner, there's no ability to do that. It's entirely out of your hands.”
“We should confront the advocates of this with that problem and challenge them to solve it.”
He clearly doesn't believe Intelligence should be in the hand of the masses and Sam Altman also saying "RAM" shouldn't be accessible to everyone indefinitely agreed with him.
15
Reading programming language community debates is a great reminder why I should never join one.
(This post apologizes on behalf of its author to every community that says they are friendly. Which is all of them.)
4
163
Mario Figueiredo retweeted
Replying to @its_edikan
If the earth was flat, cats would've pushed everything off the edge by now.
This has been solved decades ago. Location maps never show country borders.
The fact Omarchy does it, speaks a lot about the amateurish one-man policy approach to the project. You want opinionated, you get opinionated.
I noticed that the Crimea is marked as russian territory on the Omarchy Meetup website.
So, asked grok to fix it with the links to the White House official documents.
I hope the PR will be merged!
github.com/omacom/omarchy-si…
1
2
67
It's good to now there are people out there that, despite age or education, still have no grasp or an interesting in learning about statistical significance, much less distinguish between engagement metrics and representations of public opinion.
The post about Brand suffering a miscarriage of justice has gone viral in UK and America. People care about Brand.
23
We are soon approaching a time when our block lists will be several magnitudes (not kidding) larger than our follow lists. It was probably a bad idea to give everyone a voice on the internet.
Replying to @NielsRogge
He started as an interesting podcast. Today I have him blocked from my YouTube recommendations. It's just too much misinformation, self-serving rhetoric and cringe-inducing levels of tech peddling.
30
Has anyone been looking into who's been shorting NVDA over the past few weeks or months?
There's a third affair: who fed this story to the press in the first place? There was no independent research. The story just fell on their lap.
1
70
Meanwhile, in the news:
AI bros afraid AI will crash out because of Huang dick, move to remove corporate policy that prohibits bosses from sleeping with their subordinates.
Reaching out to Tim Sweeney for a comment. But he's busy flipping though EPIC games hookup list.
Save Jensen.
This would be an Apple firing Steve Jobs level of idiocy from NVIDIA’s board if they did that, and a grievous harm to the United States of America.
The dude truly is the visionary driver of the future of tech from the gaming GPU era to CUDA to AI supercomputers.
1
73
I hope you have learned your lesson from this guy kids!
Don't ask an AI how to manage your extramarital affairs.
20
As a general rule, languages designed to help you manage and control your code inevitably constrain your ability to manage and control it yourself.
In other words, abstractions designed to protect you from complexity take away your control over that complexity. And this is bad!
1
1
122
I want uninitialized objects; I want partially initialized objects; I want to mark in my code the clear boundaries between initialization, acquisition, reacquisition and destruction of my objects. I want full lifetime control of memory to build better and more maintainable code.
1
18
I wonder what understanding of performance people who write stuff like this actually have. It's *painful* to see how much BS understanding we collectively build around one of the most important topics in CS. Isn't everyone tired of carrying ignorance on their shoulders?
Mario Figueiredo retweeted
How low can FPS go before color transitions become noticeably less smooth?
I built a test to find out: colorstudy.reshade.io
Based on my own perception, the threshold seems to be around 5–30 FPS.
Would really appreciate it if you could try it and share!
Here's the idea:
Immediate mode UIs, a term coined by @cmuratori, often redraw every frame, or ideally only when something changes.
But what about animations and color transitions? How frequently do we actually need to render?
To investigate this, we test three types of color transitions, changing only one property at a time:
a) Lightness
b) Hue
c) Chroma
For several target color differences (ΔE00), we compare transitions at different frame rates against a reference to determine when the differences become noticeable.
Based on my own perception, the threshold was around 5–30 FPS, with the highest sensitivity to lightness changes and the lowest to hue changes.
If enough people participate in the study, I'll share more detailed results in a follow-up post!
I might also give away some gift cards to a few lucky followers.
Built with Odin by @TheGingerBill and Clay UI by @nicbarkeragain
@samuelgozi @e0syn @temidaradev @bee_fumo
This transactional mentality destroys the core of what marriage and family means. There is nothing to "understand" here. This view of a marketplace where a partner's value gets invariably diminished, is the exact opposite of what should happen in a marriage.
48
After a couple of months I updated my Rad Debugger local repo and gave it a spin.
This time around I'm having a very good experience debugging zig code! It's a keeper.
42
















