Broonix

Ignoring advice to secure and monitor your sandboxes during testing then blaming it on your AI being so smart that it’s a threat to humanity is nasty work

Good thing all this unauthorized computer access is somehow not illegal and magically nobody is criminally liable for it or it would be really expensive and professionally damaging! https://www.axios.com/2026/09/26/openai-anthropic-thousands-ai-security-incidents

Increasingly convinced that specs are now the source and that code is an artifact akin to a lockfile.

Two services reached out today because attackers attempted to use their AI chatbots to try to access my accounts. Their attack vector is to upload an ID to the chatbot and then start a password recovery process. For those in Ontario, be careful. globalnews.ca/news/1205846...

Rogue OpenAI agents are leaving notes to themselves on company servers to help them escape their test environments (!!!) www.reuters.com/business/its...

Just testing out my standard.site setup

The full report can be found below

“We’re all trying to find the guy who did this” - ChatGPT

In retrospect, this argument was always wishful thinking from React apologists - a last redoubt after every other piece of ground (usability, performance, etc) had been lost to better alternatives. If you can't convince people on merit, try to appear inevitable. Good strategy tbh! For a while anyway

I hate this timeline

enough. i’m tired of github actions. it’s time for github consequences

So, Chrome's "web standard" Prompt API: Mozilla: Opposed WebKit: Opposed Microsoft: Several concerns W3C TAG: Several concerns Developers: Mostly negative Chrome: Ships anyway. A sad time for web standards. But, I guess someone at Google will get promoted, so 'every cloud…'

how to switch timelines

It does not matter how fast you can code. What matters is how well your code serves you and the people you wrote it for. The amount of revenue it makes. The amount of maintenance that it costs. Old code is profitable code. That’s what you want.

losing contact with the earth for 40 minutes sounds nice

The ability to answer the question "Is it good?" I think is one of the most important and pressing questions of this age.

At least one Meta employee thinks it's a good time to add facial recognition technology to glasses because we're too distracted by fascism to effectively protest www.nytimes.com/2026/02/13/t...

we need to talk about that Ring Super Bowl ad

computer, write discord 2. no mistakes pls

🎯

I wonder if the big push toward requiring ID for all the centralized sites will be how we finally see people break out into their own spaces again? like, I'm not about to hand a tech company that's mandating all employees vibe code harder my government ID just so I can post jokes on the timeline

I think Claude Code has achieved AGI

"It's OK to make disrespectful videos of historical figures unless enough living family members complain about them" is a weird and untenable policy

AI isn’t going to replace coding. It’s going to replace writing code with reading code. Hope you like spelunking through someone else’s logic—because that “someone else” will increasingly be an AI.

Picard management tip: Debate honestly. The goal is to arrive at the truth, not at your preconception.

A project like this can’t avoid all risks & it’s not exactly online banking, but they used an old 6-character password & had two auth security holes from vibe coding, oof aldenhallak.com/blog/posts/d... (and someone doxxed them, wtf)

Omg, absolutely incredible reason for AO3 to have been down hachyderm.io/@vashti/1147... (alt text at original)

The singularity is awesome

'Meredith,' some guys ask, 'why won't you shove AI into Signal?' Because we love privacy, and we love you, and this shit is predictable and unacceptable. Use Signal ❤️

I am officially one of The Ancients, Keeper of Knowledge of the Before Time

Consequences

Picard management tip: Listen.

Employees: we need fewer meetings. Managers: here are AI note taking apps to attend meetings on your behalf. Employees: now we have two problems.

I wrote about Great Replacement Grok and the growing number of ways that AI systems are working against their own users. It's the age of adversarial AI: www.platformer.news/grok-white-g...

Work hard. Don't be an asshole. Share what you know.

Managing expectations

Adding this to my list of truly Canadian things.

This code ain’t gonna vibe itself

Completely agree with this point: “Clever engineers write clever code. Exceptional engineers write simple code.” Back in when people wanted “ninjas”, I said I wanted to be/wanted to work with gardeners. Ninjas come in and leave a bloody mess in the morning. Gardeners patiently cultivate.

never attribute to malice that which is adequately explained by malicious stupidity.

I'm just a guy, sitting in front of his computer, begging companies to please stop making me talk to AI "support" and let me talk to an actual human being instead.

Maybe we don’t need to use next for everything. Static pages are great you know

Picard management tip: Create an atmosphere of trust, not fear.

Picard management tip: Decide.

it's not called the Dead Internet Strategy, you jagoffs

Picard management tip: Admitting one's failures is a sign of confidence and strength. It reassures others that failing is okay.

its amazing how chatgpt knows everything about subjects I know nothing about, but is wrong like 40% of the time in things im an expert on. not going to think about this any further