passes Occamโs Razor with flying colors
06.02.2026 08:00 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0@loderunner.bsky.social
passes Occamโs Razor with flying colors
06.02.2026 08:00 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0A self-portrait by Henri de Toulouse-Lautrec edited to place a wrench and a screwdriver in his hands.
Self-portrait โ Henri de ToolUse-Lautrec
05.02.2026 13:50 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0AI-generated code shifts the real burden of development onto *validation*, yes.
But humans can't validate your code by reading it any more than tests can validate it by turning green.
"Is it gonna work? Is it gonna work?? IS IT???" Bub, you don't get to know til you put it in production.
Claude is built by the best-paid engineers in the world, equipped with the most advanced coding agents.
Claude also has the uptime of a 2008 self-hosted WordPress.
I have in no way been aligned with the past 20 years of French politics.
But I have to admit Iโm kind of enjoying how our country has been leveraging our natural talent at being snarky assholes on the international stage. ๐คญ
There was no mastermind or grand scheme, but there was a successful effort by wealthy men, including Epstein, to destroy a status quo that they feared was trending towards greater accountability.
03.02.2026 10:54 โ ๐ 1125 ๐ 218 ๐ฌ 12 ๐ 15The culture of impunity is so strong that these men do crimes and then *email each other at length about it*
03.02.2026 11:55 โ ๐ 54 ๐ 7 ๐ฌ 2 ๐ 0Shocker: Moltbook is insecure and left usersโ API keys exposed, allowing anyone to take over any account
www.404media.co/exposed-molt...
Person: say, i am alive. Computer: I am alive. Person: oh my god.
01.02.2026 23:48 โ ๐ 22520 ๐ 4961 ๐ฌ 82 ๐ 127The greatest trick God played on humanity is making us believe being a parent was somehow related to sex.
01.02.2026 14:44 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0New emails show Musk has been lying about his relationship with Epstein.
30.01.2026 21:30 โ ๐ 2148 ๐ 774 ๐ฌ 83 ๐ 106bsky.app/profile/lode...
30.01.2026 10:47 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0If anything, this is additional evidence that Musk has never once taken care of any one of his children.
30.01.2026 10:46 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Bilbo asking ChatGPT whether he should keep the ring. After all, why not? Why shouldn't I keep it? You're absolutely right - you found it, it's been with you a long while, and it's only natural to feel fond of something that's served you so well, especially when someone like Gandalf suddenly seems to want it for himself.
donโt let chatgpt in the shire
30.01.2026 02:30 โ ๐ 287 ๐ 53 ๐ฌ 4 ๐ 4"They program the AI" boggles my mind
29.01.2026 07:13 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0People excited with humanoid robots obviously do not share a home with a dog or a toddler.
27.01.2026 13:17 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 1Cancel or... Cancel?
27.01.2026 07:45 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0Also, fuck ICE, ICE must be disbanded and destroyed.
26.01.2026 14:12 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0More from Cory Doctorow:
The parallel with the Industrial Revolution and the Luddites is jarring.
"Textile automation wasn't just about producing more cloth โ it was about producing cheaper, worse cloth. The new machines were so easy a child could use them, because that's who was using them."
As LLM-centric automation continues to devalue the labor of development, we will have less and less negotiating power. Our white collar is turning blue.
And today is the time to start organizing and unionizing. Choose the picket line before we're sent to the food stamp line.
But the thing that bosses get wrong is: the bottleneck was never the typing.
The bottleneck is the "thinking".
It was always possible to code faster and cheaper, as long as developers didn't think it through and produced low-quality work. Developers, as a rule, stubbornly refuse to do this.
Developers are pushing work, that immediately gets judged by a supercharged word-guesser. And just like before, this feedback requires a new iteration from the developer.
Code review bots are reverse centaur automation.
Code review bots are spreading in the industry, pushing LLM-generated comments on contributions at the speed of machines.
These commentsโthough sporadically usefulโare impersonal, narrow-focused, disconnected from the work as a whole, repetitive, miss the mark...
But this is not what bosses see, as what they measure is "efficiency".
The word everywhere in Silicon Valley is "now that we've eliminated the need to write the code, the bottleneck is reviewing it".
And creating the perfect "Code review agent" is the golden goose everyone seems to be chasing.
I can do this because reviewing code has become the core skill of my senior job, and I can leverage this skill fast and accurately.
Guiding and reviewing an LLM's code is a powerful and rewarding tool in the hands of an experienced engineer.
Today, I'm doing *multiple* code reviews per hour, giving feedback to the agent along the way to achieve my goals. It's high-level work, it abstracts away the drudgery of typing, but doesn't absolve me of thinking.
26.01.2026 14:12 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0Now reviewing code is not an innate skill, and it definitely doesn't come immediately after learning to write code. You get better at it over years of experience, reviewing others code and your own, receiving reviews from other developers.
Reviewing code is a senior developer skill.
But this is a centaur's job. Bash out the code at the speed of light, painstakingly review changes, prodding the code slightly towards the acceptable iteration every time.
26.01.2026 14:12 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0It's slow. It requires reading once, twice, three times. Jumping back and forth. Cross-referencing. Building a cognitive map of the code.
It's also iterative. A review produces feedback, which informs a new iteration, that goes into review. Every iteration refines the work.
With LLM-assisted coding, the typing is done at machine-speed, producing variable-quality code in a fraction of the time and cost.
What's left to do is to review it.
Code review is a complex endeavor that requires gaining deep understanding of code, by just reading the changes.