At @spacexai, Lauren Tan (@poteto) and Roman Ugarte (@romanugarte_) are on the core team that took Grok Bot from first line of code to launch in seven weeks.
It lets you spin up AI teammates with their own identities and logins, and they just keep working in the background. Their marketing team gave a bot its own ad budget. Lauren built an entire feature with her bot in one Slack thread without ever opening Cursor.
They joined us on the @LightconePod to talk about what it’s actually like working next to bots every day, and how the hard part is shifting from doing the work to figuring out what to do and why.
03:36 — Giving Bots Real Jobs
06:49 — One Bot or a Team of Specialists
10:39 — Sharing Bots Across the Company
14:46 — Identity and Permissions for AI Coworkers
18:37 — Bot-to-Bot Collaboration and Trust
22:35 — The Workflow Behind 4,000 PRs a Month
26:00 — Teaching Agents to Write Code You Trust
29:01 — Building a Personal Software Factory
33:44 — The Future of AI Coworkers
36:47 — Rethinking the AI-Native Company
40:13 — Teaching Agents to Check Their Own Work
43:56 — How Much Code Review Is Enough?
47:33 — Codebase Guardrails That Save Tokens
52:20 — When Product Judgment Becomes the Bottleneck
56:29 — From Finance to AI Engineering
An AI agent can use your account while the app still thinks it’s you.
GrayPass (YC F26) helps companies control what people and AI agents can do inside their apps. Before payments, data exports or permission changes go through, it checks who’s acting and whether the action is allowed.
From $9K to $66K MRR in two months. Live with 13+ customers.
See GrayPass in action 👇
Backed by YC and tested by NBA star Moses Moody, our glasses tell you when your brain peaks.
Your ring tracks your heart. Your watch tracks your sleep. Nothing tracks the eyes and brain running the show.
STASIS reads your 100,000+ daily eye movements to tell you:
• When you’re at peak focus
• When your eyes start burning out
• When scrolling and blue light starts costing you sleep
6 sensors. 11mm frame. 92% accuracy. No external cameras.
Pre-orders are open. Get yours.
https://t.co/GNTDFeOe7u
Touchmark has raised a $3M seed round with participation from @inceptionfundvc, @ycombinator, @Liquid2V + @oliveur, @gabriel1, @annanay, and angels from Lovable, Fireworks, Cognition, and others.
Touchmark has served the first in history inference forward contract, processed requests for quotes on trillions of tokens, onboarded over 30 providers that listed >$30M worth of inference contracts.
In the past week we had >$2M in RFQ volume, growing week-over-week.
Inference reservations are what fuels AI companies, but these commitments are non-transferable, have no price discovery, and no unified ledger to manage them.
We are a team of IOI medalists/ex-Citadel Securities/ex-Revolut/ex-Lovable/MIT that saw this pain first hand and knows how these were solved by commodities markets that power the global economy today.
We want to solve the AI economy, and the venue for inference forwards is the first step. We approach the financialization of compute from first principles, starting with what the companies running on inference actually need, covering the full-stack - from procurement to management.
Got deals set to close by Dec 31? In 2025, 40% missed that date. Here’s why 👇
Your CRM says December. The buyer has said something drastically different over email.
Ergo flags the deal and drafts the follow-ups.
Sales teams at @retellai and @usepylon use @joinergo. We're backed by @ycombinator.
Book a call at joinergo [dot] com.
Introducing Organizely (YC F26): an AI-native ERP to help the 268,000 factories in the US lower their production costs. 🇺🇸
After running our own factory in NJ for 6 years, we're using everything we learned to build an ERP from the ground up.
More at https://t.co/5cOGxovzNg
AI is the most consequential technology of the coming years. At the inaugural YC Debate, we asked: should governments intervene substantially to slow its development?
As a side note: it was particularly exciting to film this on the eve of the most anticipated IPO of all time and release it today as Anthropic goes public.
As long as we continue to have these conversations and view our work critically, the future should go very well.
Tune in to see our debate winner.
Stop overpaying for sandboxes, GitHub runners, and GPUs
Introducing https://t.co/lXdOW2HHam, the Costco for Compute
Big cloud providers raise billions of dollars for compute and sell it with a 5x markup
We buy compute in bulk from data centers around the world and pass the savings on to you
> 30+ GPU types starting from $0.20/h
> VMs from $0.06/h
> Sandboxes cold start in 0.55s
> Ultrafast GitHub runners
> Deploy custom models on Serverless GPUs
Cheap doesn’t mean slow. We’ll run any workload on the cheapest provider with the best performance, all in a single platform.
Ask your agent to get started → https://t.co/lXdOW2HHam
Most vibe-coded dashboards have exactly one user: the person who built them. They’re stuck in one person’s Claude Code session, with the data hardcoded into an HTML file.
Today we’re announcing new Deepnote apps. Build one in Claude, and a colleague can pick it up in ChatGPT. Anyone on your team can open it with a link, and the data stays live.
As a founder, you have to reshape your company around agents. That’s why @DeepnoteHQ went open source a year ago, and why we have rebuilt our apps from scratch.
Tag Deepnote wherever you work, or keep building in our cloud with the Deepnote agent. Agents build on your team’s trusted analyses instead of starting from scratch, reusing approved dashboards, workflows, and data connections.
Every app you create is powered by a Deepnote notebook on the backend, so the data stays fresh, every version is saved, and your permissions carry across every tool.
Go have fun with them. :)
Should governments intervene to substantially slow the progress of frontier AI?
In the inaugural YC debate, Jacob Coxon, an ex-Anthropic and OpenAI researcher, faces off against Robin Hanson, an economist at GMU, about AI takeoff and what government regulation should (or shouldn’t) do about it.
00:00 - YC Debate Topic: Should We Slow Down AI?
02:42 - Jacob's case: Superintelligence by 2028
07:02 - Models already behaving in ways we don't understand
09:18 - Robin's case: the danger of abstraction
11:22 - Why new technologies take decades to diffuse
13:12 - Liability instead of regulation
14:50 - Why regulating AI is different from regulating drugs
16:00 - Jacob's rebuttal: general intelligence is one number
17:36 - Why normal liability won't cover extinction risk
19:52 - Robin's rebuttal: regulating this means regulating all compute
21:44 - AI as our descendants
23:38 - Diminishing returns
24:40 - Direct exchange: is it just the rate of change?
26:48 - How long until most cognitive work is automated?
29:21 - The jobs we have no data for
32:45 - How do you align the regulators?
35:03 - Reward hacking, and what Hugging Face actually showed
36:11 - Why not an equilibrium of AIs watching AIs
39:50 - Q&A: what regulation would you actually want?
41:36 - Q&A: what about China?
42:58 - Q&A: an FAA for AI labs?
43:45 - Q&A: the military is already deploying this
47:46 - Q&A: does regulation just move the reward hacking up a layer?
51:27 - Q&A: what unemployment rate can an economy withstand?
56:25 - Q&A: what's the downside of a pause?
58:09 - Q&A: what would change your mind?
1:02:01 - Closing statements
1:04:24 - The final audience vote
@AnthropicAI goes public today with $900bn+ committed to compute. Will it get built?
@lato_ai raised $4.7M to answer questions like this and make thorough research the standard.
In 5 days, LATO's agents interviewed 29 people across the data centre supply chain and read 261 publications. The answer is in our first public report, Ear to the Ground, now on our website (link in the comments).
LATO is a research system. Its agents find and interview the right people, read everything published on the subject, and simulate what happens next, around the clock and in any language.
Since our first study at the end of July, we’ve run 1,000+ expert interviews in 35+ countries for funds managing nearly $500bn, and passed $1M in annualised run-rate revenue in under 100 days.
The world spends more than $150 billion a year researching how people will behave. We believe agents will do that work, and LATO intends to be the company whose system runs it. We’re starting with institutional investors, and the same approach works for enterprises, scale-ups and consulting firms.
The seed round is co-led by @airstreet and @axiom_vc.
We’re also happy to be backed by @ycombinator, @inceptionfundvc, Montis VC, @smokvc, and amazing angels including the co-founders of @ElevenLabs , @vercel , @Snowflake and @RogoAI , as well as @paulg, @kirbyman01, @reecechowdhry, @annbordetsky, Tomasz Trzcinski, and @barticz
ruby2ßhell: one-byte heap overflow in Ruby to unauthenticated RCE in stock Rails, GitHub, GitLab, Mastodon, Discourse and more.
https://t.co/42RbHA1GXX
Update your Ruby version (and self-hosted software)!
I did YC in W20 and I'm back for F26. The software YC has built since then is wild.
In 2020 I was dragging SAFEs into DocuSign and placing signature boxes by hand. Now it's one click: the SAFE gets signed and wire details go out automatically. Investor CRM, investor calendar, Granola integration, live pitch feedback, all built in-house.
But the best part is warm intro paths. YC has turned its community into a real network, and some days it feels like the best (only?) way to build a startup.
Six undergrad schools (Berkeley, MIT, Harvard, Stanford, IIT, Columbia) produced about a quarter of US CEOs. The rest came from everywhere.
13% have a YC link. It's still the largest concentration of founder talent in one place; no other single place gives us this much signal. (7/8)
Omar and Goot are two of my smartest bio friends from MIT, and they're building "Theranos but it works" (finally!)
They've started 3 companies before this:
- Tome Biosciences (raised $213M)
- Sherlock Biosciences (raised $140M, created the first FDA-authorized CRISPR test)
- Proof Diagnostics (acquired by Ginkgo)
Super excited for this launch and glad I get to work with them!