trapexit (@_trapexit) header
trapexit (@_trapexit) profile picture

Author & maintainer of numerous OSS projects including mergerfs; 3DO emulation, homebrew, and misc tooling; etc. Day job involves all things AIML tech.

Austin, TX
Joined June 2009
Terminus, my 3DO homebrew goes on sale Nov 30th. 100 physical copies in first batch. My new discord has a dedicated support channel if anyone runs into issues or has questions. Join: discord.gg/GTgCruUKMg 3DO tools by @_trapexit Amazing soundtrack by @MrGamingRetro #gamedev
17
52
2
255
5,802
Another day, another disappointment. I discovered a 15+ yr old memory corruption bug in QuakeSpasm, a popular Quake source port, with AI assistance. I spent my free time preparing an issue and two PRs. The maintainer dismissed them due to AI usage and decided to copy my fix.
170
117
39
3,383
212,186
As of today @ladybirdbrowser is now a Rust project with some legacy C++ that we'll continue porting over 🤓 We've made incredible progress on concurrency in our engine since adopting Rust. I believe "CSS off the main thread" is a game-changer for browser responsiveness, and we're pushing forward on it! On the safety side, it feels great to reduce our exposure to all the classic C++ traps, both temporal and spatial. One less (but very large) thing to worry about as we enter the most hostile computing environment mankind has devised: the web...
93
80
15
2,227
71,285
trapexit retweeted
Rclone v1.76.0 is out! Highlights: • New backends: Dosya, Infomaniak kDrive • 5 new S3 providers • Lots of security fixes • rclone index command • Multi-thread SFTP uploads • GUI out of beta Announcement: forum.rclone.org/t/rclone-re…
11
1
77
2,447
I'm happy to share that the physical release of Terminus is moving along well. I finished verifying a pressed/replicated check disc on 3DO hardware, and the production job will begin next week. I'm tracking mid to end November for the release since I'm waiting to receive...
2
3
16
558
You can just be an expert
BBC: Hablaremos con nuestro experto en seguridad de BBC respecto a la vulnerabilidad de PHP. El experto en seguridad: No sabía que PHP existía hasta ayer en la mañana.
131
145
12
4,065
117,624
trapexit retweeted
Lot's of development updates! FMVs, new bikes, band music, nitro boost, better graphics and a bunch of small fixes. #indiegame
11
7
71
1,155
Stages of grief about the effect of AI on open source. Denial: - AI code is slop! Nobody serious will use this. - Open source will keep working exactly like it always has. Anger: - These AI slop PRs are ruining GitHub! - Ban AI-generated code! - If you close PRs, you're not open source anymore! Bargaining: - Okay, AI can write code, but only if a human reviews every line. - Maybe adversarial AI review will make arbitrary external code intake safe again. Depression: - Wait, if code generation is basically free, what is a contribution anymore? - Is GitHub-style open source culture just going away? Acceptance: - Open source is still open source, even if the contribution model evolves. - Good issues, test cases, benchmarks, specifications, and ideas become the scarce inputs. - Upstream may stop accepting arbitrary code while forking becomes much easier. Bonus stage! Reconstruction: - Projects redesign around maintainer-controlled agents, issue-driven development, stronger test/validation infrastructure, and new ways for contributors to earn trust.
59
50
18
475
43,095
Not to be a dick, but this seems like a paper explaining that water is wet.
China published the most uncomfortable paper on vibe coding. ETH Zurich tested 100 developers in a controlled, commercial-grade vibe coding environment to see who actually succeeds. The findings are brutal. The researchers tracked computer science achievement, written communication skills, and general cognitive reasoning. They wanted to see what actually predicts vibe coding proficiency when you never touch a line of source code yourself. Two major predictors emerged. Written communication proficiency mattered. The ability to structure thoughts and articulate intent unambiguously in text directly impacts what the AI builds. But that wasn't even the main takeaway. Computer science achievement was a massive, dominant predictor of success. Even when researchers controlled for general intelligence and reasoning skills, CS background still heavily dictated who built working software and who completely crashed. In fact, CS knowledge contributed roughly twice the unique predictive variance of writing skills alone. Why? Because vibe coding isn't about writing code. It’s about debugging logic. When an AI agent builds a complex application and quietly breaks an edge case under the hood, a non-technical user looks at the glowing UI and assumes it works. They don't know what questions to ask. They don't know what logic to challenge. They lack the mental models to recognize architectural catastrophe. You can prompt your way past syntax. You cannot prompt your way past a fundamental lack of engineering intuition. The hype told us that learning to code is dead because language is all you need. The data just proved the opposite. To truly master the vibe, you still need to understand how the machine thinks.
1
2
733
Replying to @mulaney
@mulaney Had a private screening of the new Madden movie with @AmazonMGMStudios at the iconic Cecil B. DeMille Theater. Enchanted, laughed the whole time. John Mulaney is a believably brilliant supernerd playing me — cast & script nailed every relationship’s heart! November🎬
3
10
240
I am raising my ambitions for @ladybirdbrowser. "Good enough" performance is no longer the goal. We are now aiming for *amazing* performance! 🤓
24
11
4
1,204
35,355
In each and every one of these cases, the entire incident is more or less fake. The users ask the model to do something dangerous, they give it access so that it can do the dangerous thing, and then it does it. Why is this shocking?
"we screwed up sandboxing" again during a cybersecurity eval Kimi didn't even do anything malicious with it. How about these "security" companies stop advertising their own incompetence?
12
16
5
182
9,235
better post: OpenAI doesn't know what a sandbox actually is and is currently one of the leaders in technology
OpenAI's rogue models roamed the internet for 4 days and staged a second attack dlvr.it/TTlcxn
96
121
12
3,509
164,189
Summary: there is an extremely subtle bug in the Linux kernel (not a security hole, just a bug) that was causing a rare segmentation fault in ripgrep. US AI models refused to help Daniel debugging it for “safety reasons”. Thus, he was forced to use Chinese open source models to find the problem — which he successfully tracked down given their help. A real shame that he couldn't use American models to find the problem.
21
285
9
3,447
163,682
Replying to @_trapexit
As demonstrated by the comments here, it's never the fault of Linux or anything related to its philosophy. It's always the user that expects too much. Want your audio to work? Shouldn't have chosen that distro. Your graphics drivers aren't working? Shouldn't have picked that card
2
5
196
trapexit retweeted
Seeing the Linux community response on this is classic neck beard shit. For those wondering what happened it played out like this -I woke computer from sleep - 1 of my 2 monitors didn’t boot, I then noticed it was frozen in my other monitor -tried getting to terminal/dekstop/any sign of life but couldn’t - held the power button to reboot - it then failed to boot, often spitting out errors related to KDE I spent a couple hours looking online for this error, following some guides that seemed decent but ultimately couldn’t get it to work Not everyone is a Linux expert, Windows has its issues also. It’s not a black and white issue. Thanks for joining my Ted talk
KDE plasma nuked itself and now I have no gaming computer…. I’ll be reinstalling everything now cause I’m not smart enough to fix this. But i still won’t go back to Windows
268
24
23
1,424
153,518
so LibreOffice converts pointers to integers, then puts them in strings later decodes the pointers from the strings and uses them
29
9
5
313
74,647
Share this post, if you recommend the Bill & Ted trilogy to your friends! 🎸⚡
4
106
10
250
10,554