Peter Dimitrios

Computer nerd in all the nice ways. Plays well with others.

[Peter Dimitrios @PeteDaGuru]

Thread on understanding / visualizing / generating human-understandable programs from AIML insights

[Peter Dimitrios @PeteDaGuru]

#AIML How Dangerous Is Anthropic’s Mythos AI? - think about tax loopholes or laws like code (Bruce Schneier 2026-05-14): https://t.co/C1G7QPd1Af

[Peter Dimitrios @PeteDaGuru]

Anthropic secretly installs spyware when you install Claude Desktop - https://t.co/86OIMHXY8P Native Messaging bridge com.anthropic.claude_browser_extension (Alexander Hanff 2026-04-18): https://t.co/xAFdZtScQQ

[Peter Dimitrios @PeteDaGuru]

The Modern Punchcard - Augmenting Human Intellect for the 21st Century - we are in a similar time to the Doug Englebart's 1962 manifesto Augmenting Human Intellect rather than his 1968 mother-of-all-demos and Alan Kay's Smalltalk (Tom Haymes 2026-04-21): https://t.co/ShWaelxJNW

[Peter Dimitrios @PeteDaGuru]

#ComputerNerd Clear writeup on resolving a latency issue, useful for more than Rust: Untangling Tokio and Rayon in production: From 2s latency spikes to 94ms flat - rustlang (Matheus Batista 2026-04-08): https://t.co/FysRFpf0zh

[Peter Dimitrios @PeteDaGuru]

#AIML Thinking - Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender https://t.co/gDJBqlHH39 (Steven Shaw, Gideon Namve 2026-01-11): https://t.co/ftrWh3wmir

[Peter Dimitrios @PeteDaGuru]

Wikipedia blacklists https://t.co/p05B4oWqPN, starts removing 695,000 archive links - If DDoSing a blog wasn’t bad enough, archive site also tampered with web snapshots - official stmt https://t.co/6ied9na8aM (Jon Brodkin 2026-02-20): https://t.co/dpTCdxg9h0

[Peter Dimitrios @PeteDaGuru]

Good article on NetFlix - HBO buyout - Ted Sarandos Fights to Seal Warner Bros. Discovery Deal: Netflix Co-CEO on Battling Misinformation, the Future of HBO Max and His Blood Oath for Theatrical Exhibition (Cynthia Littleton 2026-02-20): https://t.co/jUUosKkX6t

[Peter Dimitrios @PeteDaGuru]

Like playing a TARDIS - Music in the 70s - Robert Margouleff and Malcolm Cecil with Stevie Wonder playing the TONTO synthesizer - The Original New Timbral Orchestra https://t.co/0cEAqFxsuW #MOOG #SYNTH (Daniel Keller 2026-02-08): https://t.co/XKUrOoAxUG

[Peter Dimitrios @PeteDaGuru]

#AIML AI’s Wrong Answers Are Bad. Its Wrong Reasoning Is Worse - examples from Stanford KaBLE James Zou arXiv:2410.21195 and medical agents Yinghao Zhu arXiv:2510.10185 (Edd Gent 2025-12-02): https://t.co/rBRH36lHGM

[Peter Dimitrios @PeteDaGuru]

#AIML #Knowledge The Scarce Thing - thinking and teaching in the age of AI - research taste to set direction under uncertainty (Carlo Iacono 2025-12-28): https://t.co/RxExCAFQP4

[Peter Dimitrios @PeteDaGuru]

#AIML Nested Learning: The Illusion of Deep Learning Architectures - Paper: https://t.co/EyW1KCdyhX Blog: https://t.co/5TcnLcah4u - HOPE self-modifying Titans Memory System (CMS) (Ali Behrouz, Meisam Razaviyayn, Peiling Zhong, Vahab Mirrokni 2025-11-10): https://t.co/6j0F2UqlpN

[Peter Dimitrios @PeteDaGuru]

#AIML #OCR #PDF A picture is worth 1000 words but does not need 1000 tokens - DeepSeek-OCR: Contexts Optical Compression arXiv:2510.18234 - vDeepEncoder and DeepSeek3B-MoE CLIP Gundam mode -OmniDocBench (Haoran Wei, Yaofeng Sun, Yukun Li 2025-10-21): https://t.co/CWpO4nmhXn

[Peter Dimitrios @PeteDaGuru]

#ARXR #AIML Easy and Fast? Rethinking the Future of Content Creation Tools - https://t.co/7zOV7ngvLG (Takeo Igarashi 2025-10-01): https://t.co/NCLKFrENH7

[Peter Dimitrios @PeteDaGuru]

#AIML The real (economic) AI apocalypse is nigh (Cory Doctorow 2025-09-27): https://t.co/NRf3wBzC4k

[Cory Doctorow NO LONGER ON TWIT TER @doctorow]

Like you, I'm sick to the back teeth of talking about AI. Like you, I keep getting dragged into AI discussions. Unlike you‡, I spent the summer writing a book on why I'm sick of AI⹋, which @fsgbooks will publish in 2026.

‡probably

⹋"The Reverse Centaur's Guide to AI"

1/

[Peter Dimitrios @PeteDaGuru]

#AIML #ARXR "We made painting feel more like typing, when we should have made typing feel like painting" - "Putting the You back in UI" from Our interfaces have lost their senses (Amelia Wattenberger 2025-03-11): https://t.co/FYELz0jAUY

[Peter Dimitrios @PeteDaGuru]

#AIML "It is time to stop pretending that we are just one more model away from magic. We are not. The magic trick is over. The rabbit is not coming out of the hat."
The LLM In The Room: ChatGPT Speaks GPT-5 in it's own words (Jason Gorman 2025-08-08): https://t.co/U07oEH3pOG

[Peter Dimitrios @PeteDaGuru]

#AIML #ARXR some technical background:
Genie: Generative Interactive Environments arxiv:2402.15391 (Jake Bruce, Michael Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai, https://t.co/FxBtXCASJK. 2024-02-23): https://t.co/JkD8TBKq65

[Peter Dimitrios @PeteDaGuru]

#AIML #ARXR Google just casually dropping the matrix on a random Tuesday - Video - Genie 3 creating dynamic worlds in real time (2025-08-05): https://t.co/S7mrSnPlRm

[Peter Dimitrios @PeteDaGuru]

#AIML Notice how carefully they frame a 19% decrease to avoid pissing off fans of AI Assistants. Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity Joel Becker , Nate Rush , Beth Barnes, David Rein 2025-07-10): https://t.co/OZ8e0JpsnX

[METR @METR_Evals]

We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers.

The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.

[Ravid Shwartz Ziv @ziv_ravid]

You know all those arguments that LLMs think like humans? Turns out it's not true.

🧠 In our paper "From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning" we test it by checking if LLMs form concepts the same way humans do @ylecun @ChenShani2 @jurafsky