Sheldon ⚡ Hull

expert at spline reticulation. ⚡ AWS Infinidash MVP ⚡
I ❤️ coding. Work as Senior Software Engineer. My digital garden:

[ThePrimeagen @ThePrimeagen]

Within 2 days we have hit the 10,000 word Twitter article explanation of what a loop is.

This is by far the worst timeline

[CooperBaggs 💰🍞 @edgaralandough]

STRANGE FATHERHOOD ADVICE I GOT FROM AN OLDER DAD THAT ACTUALLY WORKS:

1. Never lecture your son when he's already feeling small.

[GitHub Projects Community @GithubProjects]

OpenLogi is a native, local-first alternative to Logitech Options+ written in Rust, controlling HID++ mice without an account or telemetry.

- Interactive GUI with clickable mouse diagram and action picker
- 37 built-in actions plus recorded custom keyboard shortcuts
- DPI presets, SmartShift toggle, and per-application profile overlays
- CLI for headless device inventory and diagnostics

[Sheldon ⚡ Hull @sheldon_hull]

@navanchauhan When you hit bottom make its eyes bug out and wings flap in surprise. 😆

[Mitchell Hashimoto @mitchellh]

I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.

As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!

I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.

It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?

88ms => 1.5ms
150K allocs => ~500 allocs

Incredible right? Nope.

My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.

This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.

The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.

Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.

[Daniel @growing_daniel]

Another everyday luxury is leaving your faucets running. I haven’t turned off my faucets in years. It’s convenient for when I need to use it and it creates a pleasant running water sound in my home, like a babbling brook.

[Darren Shepherd @ibuildthecloud]

I'm lost in a land of no IDE. I managed to alienate myself from all IDEs. I don't know what to do now.

[Daniel @growing_daniel]

A snake showed up in our backyard yesterday. Instead of panicking, I sent this photo to my OpenClaw. It identified it as a harmless gopher snake, gave me a local number to call to confirm, and suggested ways to catch it to make it a pet for the kids.

I have small moments like this everyday now, usually less dramatic, that remind me how AI has made daily life feel a little more magical.

[Sheldon ⚡ Hull @sheldon_hull]

@swyx Mind adding a license in there? I'd love to play with some of these, but like to stay kosher with license copy :-)

[Sheldon ⚡ Hull @sheldon_hull]

@spboyer I like.
I built something for my org as a concept that also did readability and cognitive text scoring but was Claude harness. Great minds think alike?

I looked at this yesterday and if I was using copilot more I'd be for sure running through more of it.

[Hedgie @HedgieMarkets]

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just four months. American AI software prices have jumped 20% to 37%, and GitHub (owned by Microsoft) is dropping flat-rate plans for usage-based billing across its products.

My Take
The AI subsidy era is ending in real time. The same company that put $13 billion into OpenAI and built the Azure infrastructure powering most of Anthropic's compute just looked at the bill from a competitor's coding tool and decided it was not worth paying. That is not a productivity failure on Anthropic's end. Token-based pricing is forcing every enterprise customer to confront the actual cost of running these models at scale, and the number turns out to be far higher than the flat-rate experiments suggested.

This ties directly to my Gemini Flash post yesterday. Anthropic, OpenAI, and Google all raised effective prices in the last six months. Enterprises that built workflows assuming AI costs would keep falling are now watching annual budgets evaporate in months. Two outcomes look likely from here. Either enterprises scale back AI usage to fit budgets, which slows the revenue ramp the labs need to justify their valuations ahead of IPOs, or the labs cut prices and absorb the losses, which makes the unit economics worse at exactly the wrong moment. Both paths land in the same place, the numbers stop working, and somebody has to take the writedown.

Hedgie🤗

[Sheldon ⚡ Hull @sheldon_hull]

@adocomplete @bcherny the single biggest command I use is rename. I'd like to see this be automatically done without us needing to add hooks or remember. 🙏 Swear for a period it started doing this but I have to invoke this myself now everytime I want sessions that aren't a guid or "done"

[Sheldon ⚡ Hull @sheldon_hull]

@NVIDIAGeForce #007FirstLightRTX

[Sheldon ⚡ Hull @sheldon_hull]

@josevalim For those already on enterprise usage based plans it might be a positive as officially blessing using with other tools is a change as before this was not permitted. Since those plans already pay usage costs with no subsidy it wouldn't be a cost savings anyway.

[Sheldon ⚡ Hull @sheldon_hull]

@durreadan01 The fact that this took a tweet for me to figure out is both a testament to they need better user communication for these features and a testament to some very elegant design.

[Sheldon ⚡ Hull @sheldon_hull]

@trq212 Great timing. I was doing this today and realizing how under utilized it was for reviews as I struggled with walls of markdown. Thought I was being a snowflake, but I'm not alone apparently and I'll tell myself great minds think alike. 😆

It does bring back some joy to reading.

[The Shift Journal @TheShiftJournal]

Elon Musk outlined his five-step process for applying first-principles thinking to build anything:

[Brad Wilcox @BradWilcoxIFS]

Just remember: Your coworkers will NOT be there for you in the last chapter of your life.

[Jake 🎉 @jitl]

@rough__sea try mise:

curl https://t.co/MUFjmiu4Jz | sh
mise use codex

[Zach Lloyd @zachlloydtweets]

https://t.co/cjyoRlzgxO