Diogo, Erik, and the folks I've met at TypeSafe are incredibly thoughtful (and cracked). Congrats to the whole team on the launch! Can't wait to see what's coming next
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
I've had the pleasure of working with Farhan at multiple past adventures, he's an top tier engineer and he ships like crazy.
Robotics data is a huge frontier to unlock the next wave of the Industrial Revolution, and I can't wait to see him solve it!
Today, we're introducing @shotwellst.
We're solving the hardest problem in robotics: understanding your data
There’s no point collecting robotics data if you don't understand whats happening in it, but your options are limited. If you need dense, accurate labels with precise subtask boundaries, VLMs don't cut it, an in-house annotation army is a huge operational sink, and today's vendors charge an arm and a leg for low quality work
Shotwell combines the best of both worlds by training our own annotation models and letting humans deal with the ugly edge cases.
Its the ONLY way to consistently guarantee 100% quality in a scalable way. Humans are inconsistent, and models are inaccurate.
Daniel retweeted
Offhand —
* Vacillation on masks, with abundant motivated reasoning in every case.
* Promulgation of made-up thresholds with no evidentiary basis (e.g. 6 feet).
* Authoritarian delight in nanny state intrusiveness (policing the beach and such).
* 180 on many issues around BLM.
* Lack of effective response from science funding bodies.
* Denial of aerosolized transmission.
* Changing of trial readouts so that they’d occur after the election. (Confirmed to me by senior OWS officials.)
* Crazy criteria for vaccine distribution.
* Adamant insistence on vaccine efficacy beyond what was supported by data.
* Almost complete lack of follow-through on OWS (on pan-variant vaccines).
I’m sure there are more, but those are the ones that stick out.
my friends make fun of me for glazing Elon
but it's like how could I not, the whole arc is just so insane and incredible
if you wrote a novel about a guy like this they would laugh you out of the room for being so unrealistic
towards the end of Think and Grow Rich, Hill writes about "Sex Transmutation"
it's not really talked about much bc it's kind of a touchy (hah!) subject - but TLDR it's a good idea to figure out how to channel your various energies and motivations into useful work
another way to put this -- having a crush and an enemy are both very good forces in your live to achieve big things
xitter.cf/HeidiPriebe1/status/20…
Daniel retweeted
we live in age of great moral panics about things that don’t matter at all and zero moral outrage over some of the most egregious societal sins we’ve ever seen
Unbelievably radicalizing exchange
The fact that Newsom (correctly?) thinks this will gain him points is horrific
Any competent politician (forgive the oxymoron) would tighten budgets at every turn
To openly mock those exposing the disgusting fraud and abuse ...
Insane.
Daniel retweeted
Longest MRI wait times in the developed world + the only same day euthanasia service anywhere in the world
The purpose of a system is what it does
Why does Canada have the longest MRI wait times in the developed world?
The Ontario health ministry has a central licensing department for all imaging. In 2024 less than 30 licenses were given out. I know imaging centres that have machines collecting dust for years waiting to be licensed.
This is a classic Soviet style artificial supply constraint created by bureaucracy.
Meanwhile Canadians are driving to Buffalo or overseas for private MRIs.
Daniel retweeted
If you're young and built something real instead of padding a resume -- garage projects in high school, robotics, pro gaming, sneaker bots, shipping side projects while your friends were joining college clubs, building things instead of taking internships you didn't care about -- DM me. I don't care about your GPA. I care about what you couldn't stop yourself from making.
Many talented people over the past few years were declined an offer or even an interview @xAI. My apologies.
@BarisAkis and I are going through the company interview history and reaching back out to promising candidates.
and it seems like exactly what I've been looking for
no setup, hosted for you (no Mac Mini needed). They’ve got a curated skill market — hand-picked workflows you can one-step install instead of hunting and wiring everything yourself. Inbox triage, briefings, research, content ... pick a skill, hit install, and you’re running a production-grade OpenClaw in minutes
the "production grade AI assistant" space is going to blow up, and I imagine they’ll ship full-blown AI employees soon
watch this space!
btw if you are serious about OpenClaw or similar, the best advice I have is:
use OpenRouter (easy observability), and export your API calls to LangFuse (free plan)
do a few tasks, and read the prompts in LangFuse. Notice where it gets confused. Read reasoning traces, tool calls, etc. Review the system prompts for irrelevant jank.
This is the fundamental analysis that allows you to tweak everything else, add new skills, etc. Don’t tweak your agent in the dark!
