Here's a thread of every app I've built 100% with @cursor_ai using Claude. These are all fun side projects I've worked on in my free time over the last few months.

ls works fine? Sure, but a contemporary rewrite from scratch can reevaluate some archaic approaches to how it works or displays output etc It will certainly have some issues at first, but no reason to not play around with modernizing old tools, and gradually improving them

As much as I talk shit about Rust rewrites, refreshing OSS tooling every couple decades is good actually.

Ads are a risk. Google also serves malicious sponsored results. Another reason it’s safer to browse with @brave

One interesting concept in cybersecurity, reliability engineering, and safety‑critical systems is the "dynamic non‑event": the maintained absence of failure under changing conditions. The lack of bad outcomes (like a plane crash or a breach) is not luck; it’s produced by

Absolutely love this cover. Stunning work, @_TamaraWinter @stripepress @stewartbrand.

FWIW, the conclusion I increasingly believe is that measuring the security state of an environment is somewhat like measuring human health: While there's never going to be one number or even a few numbers that are simple yet comprehensive, good metrics are still indispensable.

Constraints drive creativity. Which low/no funding, you HAVE to be creative. Abundant funds bring delusions of grandeur and overengineered solutions.

Illegal number: ps3 key (thank u @realGeorgeHotz)

Same as today but the AI morally absolves them of the burden of mass destruction

“You’re absolutely right!” *flattens residential city block*

Are there security program metrics you actually like? Other than misleading stats that lie to leadership, it’s hard to find meaningful metrics that assess program success

The most insane founders are restaurant owners: Imagine opting into the most competitive industry, with the most entitled and insufferable clients, for low profit margins.

I still have no idea what it means. “I write code sometimes”?

https://t.co/8llq9RYVdv

Traditional sast tools are in trouble: low accuracy, high false positive rate, too dumb, don’t understand context. We’ve had this problem for a while (findings piling up, no one fixing them), new tooling is promising since it helps automate triage.

What does the role “member of technical staff” mean?

What a stupid point, one is a subset. Absolute nonsense. One McDonald’s year is five thousand discovery of fire ones???

No one is handing you millions out of the kindness of their heart, no strings attached

I’m convinced investors can be detrimental to your startup, and should be avoided when possible (a luxury not many can afford)

yup, same goes for programing languages and preventing the user from shooting themselves in the foot by default