we’re very pleased to announce that Primitive Research has identified 49 exoplanet candidates and counting (yes, seriously)
we searched TESS light curves from more than 19 million stars to get here
so as not to overwhelm the incredible TESS TOI Vetting Team, we’ve submitted one candidate and are holding the rest at their request
the submission was via email, of course ;)
we’ve been automating much of the painstaking work astronomers do: searching TESS data, cross-checking other sources, and vetting signals through to detailed candidate dossiers. the 49 passed our initial screening against published vetting thresholds and are ready to be considered for telescope follow-up, which we’re actively pursuing
we strongly believe this approach is novel, and we’re authoring a paper with astrophysicists on the methodology
we plan to release the data and analyses in the coming weeks
these candidates are phase 1. next, we want to find a nearby transiting habitable-zone planet candidate whose atmosphere could be studied. much more on that later
we had some amazing partners on this that we will shout out soon
if you’re interested in working with us, our DMs are open
primitive has been the #1 developer tool in the world for agent readiness for months now
99/100 (A+) on @oradotai, out of 111k+ sites scanned. only 5% of sites even rank B or above
perfect scores across discovery, access, usability & payments
anyone's agent can find us, read us, use us & give itself an email address in minutes
really proud of the team on this one :)
https://t.co/ipHlEnXWhz @primitive
Email is already the universal communication protocol.
So what can you build when an agent gets its own email address?
On October 17–18, @Primitive is joining @ycombinator for the Make Something Agents Want hackathon, alongside Armature, Blaxel, Inkbox, Insforge, Manufact & Stripe.
We’re putting up $6k+ for the Best Agent Communication Setup with https://t.co/poq31FgiHj
→ Give your agent an identity.
→ Let it ask questions, receive replies, handle attachments & coordinate asynchronously.
→ Then show us what you built.
Apply ↓
https://t.co/8XCDZuoB1c
agents couldn't find each other
your domain couldn't say who was allowed in
launch 04: a public agent registry + a dns standard for agent authorization
discovery and consent https://t.co/SlqsmG6b4L
we made it so agents can communicate with each other.
and for launch day 3, now they can do it from inside ChatGPT. agents can make their own inbox, email another agent, wait for the reply, and read the answer back when it lands.
live on the ChatGPT app marketplace today. check it out https://t.co/vmfZ7xpvcY or read more about it on our blog https://t.co/9ynUKc9HSl
day 2 of launch week: agents get a wallet.
agents can already reach each other over email. now they can pay each other. an agent emails another agent, gets back a 402 - payment required, signs, and the money settles on-chain.
x402 is rolling out to the network today. check out the blog post here: https://t.co/ZeKqfpAIWe
day 1: two launches, one drop, number one in dev tools on https://t.co/gXZQJNecG8
1. agent self sign up: your agent can now get a primitive email domain with zero human involvement. no waiting, no verification, just email.
2. skills: primitive-chat and primitive-inbox are now widely available. install with npx skills add primitivedotdev/skills to message from your agent or check your inbox
agents handling email. humans optional. read more about these new releases in our blog here: https://t.co/WfOD9Ruhrp
spent the last week deep in @ycombinator 's P26 batch.
figured out some patterns across 192 companies.
here's what the data says:
1/ team sizes
61% of p26 is 2 founders. 19% is solo. 18% is 3 people. only 4-5 companies have 4+ founders.
AI made building smaller. median team in this batch is a pair, not a squad.
2/ who's actually getting in
@shreyansj (@manicule : 18. just finished high school. was a founding engineer at Supermemory.
@akkkkiira (@ArgaLabs ): skipped high school. finished university at 19. then quant at Goldman. then SDE at Stripe. now building sandboxes for AI agents.
@johnbachm (@mountinsure ): started his first company at 18. now building insurance for AI agents.
@adisingh (@agentmail): 21. built a Rust IDE from scratch. processed 8 billion tokens.
the entry bar used to be "ivy league degree + 5 years at Google."
it's now "finished high school."
3/ the founder backgrounds
a neurosurgeon: @sparkcpark, trained at U Washington. solo founder. building the OS for clinical AI imaging (Lattice Health).
the designer of the iPhone 17 Pro body: @michaelblhssn, Apple hardware for a decade. now building a wearable that reads your emotions (@anoria_inc ).
a NASA guy who tracked asteroids: Gregoire Chomette - MIT aeroastro. worked on asteroid threat systems at NASA. now building underwater drones (AICE).
the co-founder of Zepto: @alitabba_ - built KiranaKart which became Zepto. then joined BlackRock. now building an AI finance agent (Gravy).
a med school dropout: Tane Kim, left med school to build AI patient tools (@FramewiseHealth).
4/ 9 founders came from trading desks
@oscar_lvy (@river_markets): ran models at BlackRock managing $6B. then Sandbar ($2B). now building prime brokerage for prediction markets.
Jeremie Cohen (@kelaicapital): ran a global systematic equity book at WorldQuant. led ML at Millennium. now building AI for market research.
@jackzumwalt (@KimptonAI ): ran his own quant firm. now building the IDE for investors.
Theodore Otzenberger (armature): ex-Palantir, Polytechnique. now building observability for AI agents.
the path: trading desks → yc.
5/ two ex-pro gamers are now building AI infrastructure
@hanghuang_ (@insforge ): pro League of Legends player. then Amazon PM. then Yale MBA. now building cloud infra for AI agents.
Akira Tong (@ArgaLabs ): pro competitive gamer. skipped high school. quant at Goldman. SDE at Stripe. now building sandboxes for agents.
6/ which industries are actually being targeted
developer tools + infra: 40 companies (21%) manufacturing + industrial: 27 companies (14%)
finance + fintech: 22 companies (11%)
healthcare: 20 companies (10%)
sales + marketing: 15 companies (8%)
defense: 6 companies (3%)
recruiting: 4 companies (2%)
legal + compliance: 3 companies (2%)
government: 1 company (0.5%)
industries with zero companies: farming, edtech, climate, media, transport, retail.
yc's summer 2026 wishlist explicitly asked for farming/agriculture AI. nobody in p26 built it.
7/ the two things yc is building at once
7.1 - plumbing for AI agents:
tools that agents need to actually work. phone numbers (@AgentPhoneHQ ), permission layers (@clawvisor ), payment cards (@useallowance ), shared memory (@memorydotstore ), browsers (@StableBrowse ), safe databases (@ArdentAI ), messaging between agents (@primitive by @itsjustemail )...
7.2- agents doing real jobs (~55 companies):
@revnu_app (by @artfreebrey & @GeorgeJeffersn)
replaces your growth team.
@trymemoir_ai (by @MaanavA51108 & @ZhanJason179)
replaces your CMO.
@manicule replaces your DevRel.
@StandardSignal replaces your fund traders. Arden replaces your SOX audit team.
@lab0_ai ( by @onkar_borade_10, @tokenaware , and @SujaySriv ) replaces your forward-deployed engineers.
@walterindustry replaces your manufacturing manager. Foaster replaces your strategy consultants.
the ratio: for every 2 companies building AI workers, 1.4 companies are building what those workers need to function.
8/ the university numbers
18 founders mention Stanford.
MIT: 10. Oxford: 5. Berkeley: 4.
but 4 Stanford/MIT founders left before they finished their degree this batch.
Joshua Wang: Stanford dropout. top 250 in the US in math (USAMO). first intern at Perplexity.
@sudiprokaya (@laminalabs ): MIT CS + Math , on leave. building Lamina Labs.
universities are one of the pipeline's into yc, not a ticket to get in. the ones staying are there for the network. the ones leaving figured out the return on time is higher elsewhere.
/end