Investing @ Indeed Ventures https://teapawt.org/t.co/F4nWvcVhWI Not investment advice
San Francisco, CA
Joined March 2022
- Tweets244
- Following454
- Followers112
- Likes2K
Agility Robotics is officially a public company! They began trading under the ticker $AGLT on the Nasdaq today.
Rough day in the market to start trading but regardless, congrats to the whole @agilityrobotics team!
The Hugging Face for robotics is being built by Hugging Face !
Robotics needs open benchmarks, evals, and leaderboards. Benchmaxxing is a proven way to build better models. Not the only way, but it’s very important.
RoboColosseum (robocolosseum.ai) tests real robots on tasks like stacking bowls, opening drawers, folding towels, loading a washing machine, or carrying a box.
Tasks are performed on Unitree G1s, dual-arm YAMs, Frankas, and SO-101s, in real homes, offices, and public spaces.
Open track uses fresh on-the-spot instructions. Fine-tune track gives every model the same 100 demos, then scores 25 attempts.
Public leaderboards + all data on Hugging Face.
I’ll be posting regular updates on leaders and leaderboard changes.
Thank you to @huggingface @RoboColosseum @BitRobotNetwork @DJiafei @eaic_sg for building this!
Introducing RoboColosseum
An open robot eval initiative by @huggingface @BitRobotNetwork @eaic_sg @DJiafei
Access:
- 100+ robots (@UnitreeRobotics G1s, @LeRobotHF SO-101s, YAMs, Franka and more)
- 140+ in-the-wild scenes
- Open rollout data
- 10,000 FREE eval hrs
Thread ↓
Kyle Getsiv retweeted
4/ We evaluated 5 VLAs:
- MolmoAct2 (@allen_ai)
- GR00T N1.7 (@NVIDIA)
- π0.5 (@physical_int)
- LingBot-VLA v2 (@robbyant_brain)
- G-0.5 (@GalaxeaDynamics)
And 3 frontier models:
- GPT-6 Astra (@OpenAI)
- Opus 5.5 (@AnthropicAI)
- Grok 4.7 (@xai)
We tested them across different robot embodiments, with evaluators manually judging the results.
Blog & paper: robocolosseum.ai/blog
Leaderboard: robocolosseum.ai/leaderboard
Uber is not asleep at the wheel in resturants & delivery!
Q3 delivery volume:
• DoorDash: $33.9B, +23% (excluding Deliveroo)
• Uber: $28.5B, +23%
$UBER CEO: "We are very excited about the progress of our grocery and retail business. We are run rate now at a $15 billion run rate, growing 40% on a year-on-year basis. We think this is a $10 trillion TAM, so it is an enormous market for us"
Heading to @corl_conf in Austin next week!
A few things I’m most interested in:
- Real challenges of deploying robots (last-mile reliability, failure recovery, hardware variance)
- VLAs, World Models / World Action Models, and what Astra actually changes for the robotics stack
- Is the bottleneck still data scale & quality, or are we seeing meaningful architectural unlocks?
- Sim2Real that holds up outside the lab
Very excited for the keynotes from @drfeifei and @RussTedrake, plus the panelists @chelseabfinn @pulkitology. Also looking forward to @DrJimFan’s talk on World Action Models and agentic robotics.
DM if you’re around and want to chat! I’ve spent a lot of time on humanoid unit economics and fleet financing.
Got off the waitlist today and have loved it! Favorite persistent agent so far.
Instinct didn't really stick for me, not a fan of the "no app" approach, I want a good UI where I can see my chats, my agents, etc.
Muse has been good and knows a lot about my personal life through instagram, but it's not the complete platform that Claude or Chat is.
Claude already has my all of projects and context. It's connected to my email and is suggesting which emails I should respond to.
Most importantly, Duet lets me interact with the local files on my laptop (my local files are synced to Google Drive) so I can actually do long-running projects without leaving my laptop open or buying a mac mini.
Kyle Getsiv retweeted
Today we’re out of stealth, and we’ve raised a $40m seed round co-led by @creandum and @usv, to bring Instant Robot Learning to life.
For the past few months, we‘ve been busy building a new company.
Machenta.
This follows results from a new kind of robot foundation model we’ve built, which I’m so excited to share with you today!
Usually, it takes hours/days of data collection and training to teach a robot.
We’ve built a model that learns tasks instantly: one demonstration, no further training.
A model trained entirely in simulation, without a single pixel of real-world data.
A model where teaching robots is so quick and easy, it feels like physical vibe coding.
A model for robots that learn like humans do, from someone like you.
Welcome to the era of Instant Robot Learning.
Thank you @Kathi_Wilhelm, @rebeccakaden, @hkanji, @jonno_evans and all of our angels for your support!
@Machenta_
Learn more: machenta.com/blog/instant-ro…
See below for what you can do with Instant Robot Learning!
Kyle Getsiv retweeted
You've been grinding on S-1 teardowns for years. Dozens of boring SaaS companies. The time is now. This is your moment.
The best enterprise SaaS companies are building vertical AI harnesses that combine their system of record with agentic workflows so they can own the end-to-end job. Horizontal agents and harnesses aren't designed to dominate individual verticals or departments, and they won't build the depth, domain-specific data, or learning loops needed for strong interactions in their areas of expertise.
Ramp Glass will always be better for FP&A than Codex or Claude with a Ramp connector, and they can go multi-model to save on cost. @harvey and @WeAreLegora are doing this for legal, @RogoAI for finance and capital markets, @tryramp for FP&A, and @figma for design.
I'm not sure OpenAI and Anthropic will pull off the mass enterprise deployments they expect. Salesforce is much larger than Microsoft Dynamics and Workday is larger than Microsoft Viva. The frontier labs can still sell general licenses and make a lot of money on token consumption, but they may need to stay focused on core horizontal workflows. Deep vertical work in legal, finance, and similar areas requires the compliance plumbing, data-room and systems-of-record integrations, change management, and customer relationships that the labs have little incentive to own. @GabeStengel and @winstonweinberg have said this directly: the durable edge is building perpendicular to the model layer.
Congrats on the Pi API beta! Looks like some really cool partners are already using it.
Can’t wait to see what people build with this.
Kyle Getsiv retweeted
Assistant Benchmark, a personal agent benchmarking platform, is already tracking 139 agents, across horizontal and vertical specific (travel/shopping).
Currently has Muse as number 1
Q3 book reviews!
Apple in China, @PatrickMcGee_
The best account on how Apple built its legendary suppy chain and subsequent China dependence. Apple was far more hands-on than I knew. First book I've read that covers China in depth, definitely need to read more.
1873, Liaquat Ahamed
So interesting to read about how a couple of wars and capital-intensive buildouts led to the first global recession. It also covers the growing pains of globalization and credit markets.
Empires of Light, Jill Jonnes
The War of the Currents. Edison, Tesla, and Westinghouse built modern electricity, and none of them kept control of their companies.
Investing in U.S. Financial History, Mark J. Higgins
Such a great book on the economic history of the US; I keep it on my bedside and re-read sections before and after period-specific books.
Open to any and all book suggestions!
Excited for CVC week!
If you’re attending, drop a comment or send a DM. Let’s find time for coffee or catch each other between sessions
#CVCWeek #CounterVII #CorporateVenture
CVC Week is back! 30+ events, 1500+ investors in town. Wednesday is Counter VII the biggest CVC conference of the year. Look fwd to hearing from this amazing lineup of speakers. #CVCWeek
Kyle Getsiv retweeted
We instituted an official AI Writing Policy at @Clay.
Massive thanks to @sophiebits on our engineering team who wrote this.
Originally this was just for eng, but other teams found it so helpful we expanded it company-wide.
Here are the four guiding principles:
1. You must stand behind every idea and sentence
It is your responsibility to make sure that the entire document is representative of your own thoughts before you share it.
2. Writing is thinking
Spending time on the writing process teaches you more about your topic. If you circumvent this process, you will walk away with a poorer understanding of the subject matter.
3. More time should be spent writing a document than consuming it
If you generate a document from a short prompt then ask your readers to go through the longer output, you are disrespecting their time. They can talk to ChatGPT themselves if they want to.
4. Longer is not better
AI makes it easy to generate long docs, and it loves padding them with sentences that say nothing. If you're producing docs from a short prompt, consider just sharing the prompt.
You can read (and borrow) the full policy below. Hopefully we all communicate a bit more clearly now :)
Kyle Getsiv retweeted
Introducing Midcentury.
We’re building the data and simulation infra for physical AI.
Today, we’re coming out of stealth with a $15M Series Seed to scale robotics beyond polished demos.
We’re already supporting frontier labs with:
→ The world’s largest egocentric dataset: 2M+ hours, 50+ environments, 20,000+ tasks
→ Matrix: a frontier simulation platform scaled with our real-world data to evaluate and post-train policies at scale
This is a great concept if it’s executed with real teachers and continuity. The names involved are so impressive but most of them have full-time jobs (some have multiple jobs).
Short courses and office hours aren’t a substitute for teachers who see students every day and are actually invested in them. High-agency 18-year-olds still need that.
Still very cool!
Today, we’re launching an ambitious new school called The Horowitz Andreessen Academy.
Based in San Francisco, The Academy serves the most promising young high school graduates.
We think this can be an elite institution that attracts top tier talent. One that prepares students for the future rather than remaining stuck in the past.
The #1 goal is to help students learn to build, which is the most important skill in the AI era. They'll learn primarily by pursuing their own projects, either individually or in groups. There are classes and guest lectures, too, from some truly amazing people who have built modern-day Silicon Valley.
The Academy is designed as a network, since that’s the reason students go to school in the first place. Core to that network are our 10 Founding Partners: Anduril, Anthropic, Coinbase, Google, Meta, NVIDIA, OpenAI, Palantir, Replit, and Stripe. The network includes over 50 hiring partners and over 200 speakers and mentors. To join as a hiring partner or faculty member, you can apply on our website.
We raised $42M in funding led by @a16z. I'll be CEO and @pmarca and @eriktorenberg will join me on the board.
Applications are open for our Founding Class Fellowship, which will be one year and tuition-free. Eventually, pending regulatory approval, we plan to offer a two-year program that charges tuition, similar in cost to an elite private university.
We're looking for the most unusually ambitious young builders on the planet. Come join us in San Francisco:
theacademysf.com/
Excited to be co-hosting a mixer on AI, HR & the Future of Work for #SFTechWeek with @SHRM Labs!
📅 Thursday, October 8th at 4:30pm
📍 San Francisco
The event will feature pitches from emerging startups at the intersection of AI, HR & the Future of Work, followed by open networking.
Join us for a casual evening with free food and drinks, and opportunities to connect with founders, investors, CVC teams, and executives from @indeed and @Glassdoor, as well as the Indeed Ventures Team.
🙌 We’d love for you (and your teams) to join us!
👉 RSVP here partiful.com/e/gPHSsw6QRMO0R…