Anthropic's playbook is basically to overstate the capabilities of your model, apply domain restrictions, and sell all the fears that come with allowing it.
Midwits then think you've built a god or some form of AGI.
Ride the hype.
Rinse and repeat.
Fwiw @MIRIBerkeley has never really been on PauseAI US’s or my side, even when we did free, one-sided promotion of If Anyone Builds It. Nate has been helpful at times. Eliezer obviously felt like he should help but couldn’t bring himself to bc of how I challenged him.
People from MIRI have scolded me for being confrontational on Twitter many times, including Alex once hijacking a meeting that was supposed to be about lobby info sharing to do so. I have never publicly taken issue with their PR decisions, figuring that what’s good for the goose is good for the gander when we have a similar message.
But these are their employees (?) and leaders, being the dumbest shits imaginable 🤷♀️ They are declaring that they cannot be worked with respectably— that is the point. I guess they felt like shills doing the book tour thing and wanted to get back to their roots. Fine. The relationship was of very little benefit, I endured constant insults and lack of appreciation to be there, and now it’s become a liability. We’re out.
FYI.
sorry how exactly did u guys think this was going to go? u thought Anthropic was going to build the infinity machine that can cure all diseases and prevent aging... and then let fucking Eli Lilly extract that and get the patent?
*the labs are going to do all of it*
hi, i’m sam! 👋
i’m new to sf so figured i’d do a quick intro.
recently moved here from nyc and currently an AI engineer building in stealth.
a few things about me:
- i take my coffee embarrassingly seriously
- into pilates, padel and fashion
- i write a tiny health & wellness Substack
- a few months into learning how to dj
- always planning my next trip
if you’re into building agents, coffee, music or just looking to meet new people in sf let’s connect :)
you can design hardware in codex / claude code
this is a 7dof robot arm with functional kinematics that was 100% prompted (no cad software)
using custom skills, agents can generate STEP files, robot description files, and even run FEA checks from text / image prompts
China's Unitree Will Dominate Global Robotics
The Fastest Iteration Cycle In Next-Gen Robotics Should See Unprecedented Acceleration
https://t.co/0HKSayvvTK
The thing everyone calling LLMs 'intelligent' is missing: they're interpolation engines with trillion-dollar marketing budgets. A dog generalizes from a single experience. A $10,000,000,000 model trained on sorting packages cannot pack a car.
@Benthamsbulldog This is probably the most ridiculous attempt at arguing in favor of anything that I have ever seen. It claims there is no serious case against lab-grown meat, but then goes on to discuss what I would call superficial concerns at best about lab-grown meat.