Sunfield

Shifting to posting more about building in tech / AI research (still doing smash labbing on the side) Like always, only gonna post shit I think is cool rather than spamming. Thanks to my friends in the smash community for sticking with me the last few years 🫡🫡 face reveal(?)

I do not want to lose to R.O.B anymore.

@Tetraslam holy shit man feel better soon

today i got hit by a car zuckerberg sf general hospital is pretty goated! anyways, no lasting damage (ct scan was all negative phew) but my face will take a month to (fully) heal so in the meantime I apologize for looking like literal roadkill

“One of the most amusing things about my infrequent visits to the Bay Area is how every single person has the same heterodox takes at the same time.” - @JerusalemDemsas

You’re welcome northeastern students I willed this into existence

We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.

@taalas_inc the demo feels fucking insane this is sick

24 dedicated people. $30M spent on development. Extreme specialization, speed, and power efficiency. Today we launch Taalas’ first product. Check it out: Details: taalas.com/the-path-to-ub… Demo chatbot: chatjimmy.ai API: taalas.com/api-request-fo…

@growing_daniel reads like an incredibly good satirical fiction

This article is so good harpers.org/archive/2026/0…

@danieljh0n @darmVL @CasualScrubVAL @orca_val @PashaVLR @Maxyvalo @CasualScrubVAL first social outing ever 😮

Rutgers University’s Valorant team!

just finished reading the Sequences. so cool that I'm infallible now

Zamn

petition for @claudeai to give students access to claude code with our enterprise accounts (its literally economical to get students hooked on claude code anyways)

@xydotdot The hard truth here is that LLMs are not humanlike, humans are LLMlike

Moltbook is nothing more than a puppeted multi-agent LLM loop. Each “agent” is just next-token prediction shaped by human-defined prompts, curated context, routing rules, and sampling knobs. There is no endogenous goals. There is no self-directed intent. What looks like autonomous interaction is recursive prompting: one model’s output becomes another model’s input, repeated. Controversial outputs aren’t “beliefs,” they’re the model generating high-engagement extremes it learned from the internet, because the system rewards that behavior.

Watch this video before spending $700 + 5 hours setting up a Mac Mini for Moltbot [Clawdbot] You can do all the same things that Moltbot can but from iMessage with Zo Computer for FREE, here's how 0:00 - iMessage -> Zo buys stuff on Amazon 0:55 - Why not to use Moltbot 1:42 - What is Zo and why is it better 3:15 - Setting up Zo 4:35 - Scheduled Tasks 7:00 - Skills 8:30 - Zo's Browser (iMessage -> LinkedIn/Amazon) 10:30 - Vibe Coding with Zo 11:23 - Zo buys me toilet paper from Amazon with iMessage 13:30 - Why is something like Zo better than Clawdbot?

how to have the power of clawdbot by doing literally nothing @zocomputer has another computer built in with a bunch of automations already set up for u. zo is ur computer companion!

Peabnut is a beast how do you make mega man that hype bro

this is literally us zo is the cloud computer that can be ur ai companion

playing against this guy is the scariest shit on earth it does not remotely surprise me that he beat leo