Pinned Tweet
6+ years in Web3. I came for the innovation, stayed for the community, and learned more than I expected.
But I no longer find myself there professionally, and I'd rather be honest about it than stay for the wrong reasons.
Next chapter: Solutions Engineer at a YC-backed company, AI + mobility.
I know "pivot to AI" is a meme. But for those of us who genuinely love tech, it's also very real.
Keeping this account, but the vibe might shift a bit. Still code, still open-source, probably more books, more games, more of whatever I actually enjoy.
Thanks to everyone who was part of this chapter 🙏
The most annoying thing about going back to corporate life is having to use Microsoft Teams. Everything else is perfect.
Calling AI a "fascist-backed technology" is exactly why Linux is going nowhere without @dhh's Omarchy project.
Embrace techno-optimism or stay irrelevant, that's the current paradigm.
“F**k that, I’m out,” says prominent Debian developer, who is leaving the project after Debian Members voted to allow AI contributions.
From Antoine Le Gonidec’s resignation:
“I want the consequences of Debian fully embracing fascist-backed technology to be known.”
“I am no longer interested in anything coming from Debian.”
“Pretending to have a neutral stance when faced with fascism is not neutrality, it’s active collaboration.”
A classic example of a trendy app that shows just how little people care about privacy.
Instead of being horrified by years of tracking, they have fun watching the map come to life.
Cool. Now go turn off your Google Timeline.
Or better, ditch Google entirely.
구글 타임라인 1년간 행적 동영상으로 만들어주는 앱인데, 꽤나 만족스럽게 잘 만들어지네요. 작년에 안티그래비티가 초안 만들었던 걸 오늘 코덱스가 완성해줌. 안드로이드 쓰는 트친님들 apk 파일 받아서 써보시죠
github.com/mahlernim/google-…
A simple website to remind you that what you take for granted today was the ultimate dream of yesterday's civilization.
ordinaryabundance.com/
DeepSeek is so much more malicious than Sonnet, it's actually kind of scary.
Testing the moral limits of these models is definitely an experience I recommend.
Why do LLMs struggle so much with time management?
They can't tell what time/day it is, when a prompt was sent, and so on. But doesn't that seem like factual metadata to add to the payload?
Like, Claude literally replies "good night" to my first message every day 🥴
Approval drainer vs MEV bot 🫠
🚨Community Alert:
Blockaid Exploit Detection system detected an exploit involving the @jaredsmev MEV bot on Ethereum.
The incident resulted from attacker-controlled contracts tricking an automated MEV execution system into granting token approvals, later used to drain funds.
7.5M$ drained.
Exploit tx:
etherscan.io/tx/0x2be8704f5a…
More details in🧵
marin retweeted
If you use Gmail, you need to read this.
Google’s AI, Gemini, can read your emails, attachments, bank statements, etc.
For many, this feature was switched on without consent.
Here's how to turn it off 🧵
1/
marin retweeted
you are not a free individual if someone can take away your AI access
local models are the way.
few understand.
