I’m scared about the future of security
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Related videos
I'm scared to make this video
Theo - t3․gg
187.3k views
I’m done.
Theo - t3․gg
185.8k views
Prime is (mostly) right about AI
Theo - t3․gg
136.4k views
Claude Mythos and the end of software
Theo - t3․gg
287.6k views
I’m serious.
Theo - t3․gg
95.7k views
I'm so f***ing tired of Obsidian.
Theo - t3․gg
113.3k views
The Future of TypeScript
Theo - t3․gg
152.4k views
We need to talk about Sonnet 4.6
Theo - t3․gg
114.3k views
Fumble of the decade
Theo - t3․gg
80.5k views
I asked Sam Altman about the future of code
Theo - t3․gg
71.2k views
Top Comments (10)
I heard that obsidian is good for security
As a cyber security professional, the main takeaway I get from this is, the job hasn't changed, the tooling has. Agents are replacing manual tasks, to achieve the same end goal, the human lead tasks, such as educating employees on being safe, architecting systems to be secure by default (yes with vibe secured code, but still human managed) are going to be the future I think
After garry tan, Theodora will be the next victim of ai psychosis
"I guess technically speaking I am fear mongering" - yes, you are
I think as security researchers at companies continue using AI to look for vulnerabilities in their entire stack, there will continue to be a downstream effect where those researchers start flagging those vulnerabilities for their vendors and start opening PRs in open source repositories, and the end result is that the internet becomes more secure, not less.
Defcon? The NSA recruitment event?
Seems like they invented generative fuzzing
I just finished the finale of the HBO series Silicon Valley. I don't think I could have picked a more relevant time to watch the show for the first time.
I tried the puzzle with GPT-5.4 Pro (fail), Claude 4.6 (fail), Grok (fail lol), Meta's new model (fail), GLM-5.1 (fail), and Gemini 3.1 Pro (nailed it!). Then I got Gemini to give the others a clue: "Take a close look at the labels on those 12 bottles of rum and pair them with the second line of the shanty. If you were to take the letters on those labels and move them 'three swigs ahead,' where in the alphabet might you land?" GPT then got it, the rest still had no idea.
post-attention-scarcity in security makes me think of a zombie apocalypse where the zombies aren't easily distracted and they remember how to use tools.
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (10)
I heard that obsidian is good for security
As a cyber security professional, the main takeaway I get from this is, the job hasn't changed, the tooling has. Agents are replacing manual tasks, to achieve the same end goal, the human lead tasks, such as educating employees on being safe, architecting systems to be secure by default (yes with vibe secured code, but still human managed) are going to be the future I think
After garry tan, Theodora will be the next victim of ai psychosis
"I guess technically speaking I am fear mongering" - yes, you are
I think as security researchers at companies continue using AI to look for vulnerabilities in their entire stack, there will continue to be a downstream effect where those researchers start flagging those vulnerabilities for their vendors and start opening PRs in open source repositories, and the end result is that the internet becomes more secure, not less.
Defcon? The NSA recruitment event?
Seems like they invented generative fuzzing
I just finished the finale of the HBO series Silicon Valley. I don't think I could have picked a more relevant time to watch the show for the first time.
I tried the puzzle with GPT-5.4 Pro (fail), Claude 4.6 (fail), Grok (fail lol), Meta's new model (fail), GLM-5.1 (fail), and Gemini 3.1 Pro (nailed it!). Then I got Gemini to give the others a clue: "Take a close look at the labels on those 12 bottles of rum and pair them with the second line of the shanty. If you were to take the letters on those labels and move them 'three swigs ahead,' where in the alphabet might you land?" GPT then got it, the rest still had no idea.
post-attention-scarcity in security makes me think of a zombie apocalypse where the zombies aren't easily distracted and they remember how to use tools.