Building humanoids to turn physical jobs into remote work
NYC
Joined March 2018
- Tweets338
- Following967
- Followers864
- Likes825
My thoughts are with the French policemen trying to maintain order. At that point, Germany should offer to send reinforcements as a sign of our friendship @_FriedrichMerz
⚠️🔴 À REGARDER JUSQU'À LA FIN
JE NE PEUX PAS DÉCRIRE CE QUE JE VIENS DE VIVRE À MARSEILLE.
#17octobre
Leander Maerkisch retweeted
At this point the whole EU Inc negotiations are so lost in the details that everyone forgot the main goal: Creating a standard.
Basically, allow founders to raise quickly globally and use stock options to hire competitively.
Let’s say it quite clearly:
1) If EU Inc isn’t providing a standard it’s useless. We already got 30+ legal entities in Europe. Experienced founders use Delaware Inc already because of this.
2) If EU Inc doesn’t fix stock options Delaware Inc is still the better option.
I understand the there are complex details and people are worried about loopholes. But you can’t lose the actual goal of it because of it.
An EU Inc that doesn't provide a standard is pointless. Nobody will care about intricacies of compromises reached in political committees if the market can't use the outcome.
If you are in contact with your political leaders (prime ministers, presidents, et al) please reach out to them. Or to founders/investors in your network who might have that access.
🔥🇪🇺✌️
xitter.cf/euinc_petition/status/…
Reminds me of Rodney Brooks essay 'Why Today's Humanoids Won't Learn Dexterity'. Touch is necessary modality for robotics that is not solved yet. We need more accurate, reliable, and lower cost touch sensors.
Said hi to two of our wonderful angels in Munich, Jens & Nikolay. Would have been fun to bring the robot to Oktoberfest but unfortunately no time. Next time!
There might be market adjustment going on with the Chinese robotics industry. Apparently the Chinese govt has stopped allowing robotics companies from IPOing. Unitree stocks down from 1100yuan to 400yuan. Too much supply, not enough (authentic) demand.
Leander Maerkisch retweeted
Pivot to robotics. Build the hard thing!
for a startup founder right now the paradox is that this may be the best & worst time to start a company.
the opportunity surface is basically infinite cuz ai keeps collapsing the cost of building, experimenting, & entering new markets. but the durability of any individual advantage is collapsing faster than ever too. anything obviously valuable gets copied, bundled, or commoditized incredibly fast. before you can even blink.
& worse now the big guys are shipping fast too so in many ways you’re kind of running downhill while meta, openai, anthropic, etc. are driving behind you even faster in a giant luxury bus with the high beams turned on to the max.
Co-hosting Berlin robotics meet-up next Tue (22nd Sept), 19h in Mitte. Bringing one of our latest robot versions. For robotics builders and curious minds. Link below or DM luma.com/fy6sm639
Legendary video. Love how Diogo changes outfits every cut.
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
Leander Maerkisch retweeted
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 talk to my AI all day using FluidVoice (open-source alternative to Wispr Flow).
Last night I shared a room with my teammate Anton. At 4am he heard me giving my AI instructions.
In my sleep.
AI psychosis or productivity hack?
VCs can gain a competitive advantage by reliably giving portfolio companies early access to frontier models like Astra. Curious why nobody does it yet.
It's insane how much negative value the EU commission creates. Completely out of touch with reality, or doing anything useful.
We have designated ChatGPT as a Very Large Online Search Engine and Reddit and Roblox as Very Large Online Platforms under the Digital Services Act.
They now have four months to comply with additional DSA obligations.
More: link.europa.eu/8bmk3M
#DSA
ALT “Graphic with a purple background and yellow EU stars. Text reads: ‘ChatGPT, Reddit and Roblox designated under the Digital Services Act.’ The European Commission logo appears in the bottom right.”
Watch out for these guys. Letting an AI design a rocket from scratch and actually building it is nuts.
Reinforcement learning rewards convergence, and convergence is the opposite of discovery.
We built the world’s first judgment model from the decisions papers erase: what to pursue, challenge, abandon, and revisit. Columbus-1 autonomously found a zero-click Bluetooth RCE and designed a 10-foot, self-landing rocket.
Leander Maerkisch retweeted
Reinforcement learning rewards convergence, and convergence is the opposite of discovery.
We built the world’s first judgment model from the decisions papers erase: what to pursue, challenge, abandon, and revisit. Columbus-1 autonomously found a zero-click Bluetooth RCE and designed a 10-foot, self-landing rocket.
I've been thinking quite a bit about what it would take for a humanoid to pick up a big box w/ unknown content (=weight). Humans can do this because they predict the weight reasonably well given visual information. When lifting, our tendons measure actual mechanical load against the prediction and adjust force (feedback).
Autonomous humanoids learn by mimicking human behaviour (compression of observational data). I think the only way for humanoids to learn this generally is via teleoperated humanoids deployed in real-world settings. It's 1:1 mapping and includes human judgement under uncertainty. The alternative of equipping humans with all kinds of sensors as they go about their lives, yet this fails to measure what's happening inside the human (e.g., mechanical load in tendons).
Now, the challenging part is how to sense, transmit, and feedback the right information for a human remote-controlling a humanoid to apply their built-in judgement in a low-cost way. That's still an open question for us. Perhaps anyone has a good idea for this.