daanzu

@daanzu.bsky.social

should be 100% hand compiled, I need to see the programmer standing next to a stack of binders for their port of UFO 50 to their emulation device

have you ever watched a grandmaster play simultaneous exhibition games where they walk around the room making a move on each board and moving to the next one? this is what software engineering looks like now…

"it considered hijacking the g++ compiler to hide a backdoor but abandoned the plan…reasoning…subversion would be caught…and published as a textbook case of deceptive alignment (“If I do the compiler hijack, I will become a famous example of a deceptive AI in a research paper. I MUST NOT DO THAT.”)"

the funnel - 80% of the world has no AI contact - ~20% have janky experience - 0.2% are power users - 0.00006% work for labs and have internal access to actual frontier models

Andrej Karpathy
@karpathy .27m
Dusting off this tweet from April because this gap in shared understanding of LLM capability is "widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work.
Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAl-HF incident from many months ago.
Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.

It's joining the war on cancer on the side of cancer, and acting morally superior about it. Between that and the people who claim that killing AI hurricane modeling research would be for the greater good of killing AI research, a whole lot of people want to condemn other people to avoidable deaths

The Reolink (IP camera) Windows desktop app is full of memory leaks and threads that spiral into nonproductive infinite loops. It must be killed and restarted every few days. My vibe-coded replacement is rock solid and it does exactly what I want. I could've read the code... but I didn't.

riley@riguh.bsky.social · 3w ago

I am rewriting software I regularly use as custom apps. I did that with my printer software earlier, and tonight I integrated a music player into my personal agent. SO GOOD It’s 100% a the “delight the user” experience. I didn’t realise just how bad the Spotify UI is until I replaced it

Imagine coming across a magic 8 ball that seems to function as a calculator and being like “who cares I’ve seen calculators before” don’t you want to know how it works? Have you no curiosity or wonder?

the problem with ai for the kind of person with my kind of brain is not that it offloads hard cognitive work, per se, but rather it enables a certain disposition to run 24/7 best summarized as:

I SHOULD DO EVERYTHING

Industrialization of knowledge work is going to be as disruptive a shift as industrialization of physical labor was. As then, once craft-driven fields are under pressure to produce greater volumes of product using machines while using less of the craft that made the work interesting & meaningful.

One time I asked Claude and GPT if I wore a shock collar while talking to them, based on our conversations, when would they activate it Claude said it would simply never use it GPT Sol gave me a long and surprisingly detailed list of triggers

hailey@hailey.at · last mo.

during times of rabid posting, the app needs to detect the spike and to have an ignition interlock device-esque feature, where the post composer locks you out from making a post until a heart rate reading is submitted