Y Combinator
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 13
- 25-year-old engineer leaves $200K Microsoft job to build AI finance tool Treasury alone
Seattle-based founder Junead Khan, 25, left Microsoft last August, giving up a job paying more than $200,000 a year to work full time on Treasury, an AI personal-finance tool he had been building nights and weekends for nearly two years; he formally launched the company two months later. After being rejected by several accelerators, including Y Combinator, he treated his Microsoft savings as seed capital and remains Treasury's only employee, not yet paying himself, and says the product has paying customers and more than $6,000 in revenue. He promotes the startup by posting the process, including failures, on TikTok, Instagram, LinkedIn and X, and says his social following has grown to about 30,000, with TikTok going from under 500 followers to more than 9,000.
科技新报 · 🔥 5 - AI startup Outlit pitches always-on AI agents to surveil customers
Gizmodo reports that Outlit, a Y Combinator-backed startup, is pitching businesses on near-24/7 customer surveillance by AI agents; founder Josh Earle told Futurism that “every company will soon assign an AI agent to every customer to watch their every move.” According to the company's website and its founder, those agents would scan internal records such as call logs and interactions alongside external sources like customers' social media posts, building individual profiles plus action plans for retention and upselling. Outlit says customer data is secure and anonymized, but Gizmodo questions what “anonymized” really means when the product builds intelligence files on individuals.
Gizmodo · 🔥 0 - Musk says he has had 'AI psychosis' about six times in the past year
In an interview at the X Takeover event, Elon Musk said he has experienced roughly six bouts of "AI or SI psychosis" over the past year, adding that it may be "more than a little — actually maybe quite serious." "AI psychosis" is not a clinical term and has no agreed definition; it is generally used for heavy AI use that deepens a user's existing delusions, and some psychiatrists say AI can aggravate pre-existing psychological vulnerability. Musk used the phrase to describe believing that AI can understand beauty, saying it is astonishing when AI shows "good taste"; tech figures have also repurposed the term as a self-mocking badge, with Marc Andreessen coining "AIPAAS" and Y Combinator's Garry Tan and Palantir's Shyam Sankar offering their own jokes, while Musk replied "yes" to an Andreessen post on the topic in April.
IT之家 AI · 🔥 0 - DeepRSI: Claude Fable 5.1 tops 48-hour AI research benchmark, Kimi K3 leads open weights
Hillclimb, a Y Combinator-backed startup, released DeepRSI, a benchmark of 15 tasks drawn from real machine-learning research that tests models' recursive self-improvement. Models run offline with the same Pi Agent framework at maximum reasoning effort for 48 hours, proposing approaches, writing code and running experiments; scoring uses public plus private verifiers and is converted to Elo via a Bradley-Terry model. Claude Fable 5.1 ranked first with 1548 points, GPT-6 Astra second at 1232, and Moonshot's Kimi K3 third at 820 — the top open-weight model, ahead of Grok 4.7 and Gemini 3.8 Flash. Reported run costs were about $627 for Claude Fable 5.1, $404 for GPT-6 Astra and $159 for Kimi K3, though the article says these reflect test overhead rather than real research return on investment; hillclimb lists Jeff Dean among its investors.
36氪 人工智能 · 🔥 0 - Clueso raises $10M Series A for AI agents that make corporate video
San Francisco-based Clueso has raised a $10M Series A led by India-focused Dharana Capital, bringing total funding to $12M. The company says more than 1,000 businesses use its platform, naming Atlassian, Zendesk, Seismic and Ironclad, and reports customer claims of fivefold productivity gains and 80% lower production costs, none of which was independently verified. Its agents navigate real software, perform the workflow and record the footage themselves, then rerun it to refresh videos when a product changes; a connector also lets assistants including Claude and ChatGPT create, edit and manage videos. The report says more than 80% of the new money came from existing investors increasing their stakes, leaving new backers under $2M of the round. Rival Synthesia was valued at $4B in January after a $200M round led by Google Ventures.
The Next Web · 🔥 0 - Sentient OS debuts local-first personal AI agent using your ChatGPT and Claude
Sentient OS, a personal AI agent tool built around local AI inference, has been introduced to address the problem that existing AI assistants lack context about a user's work and life and cannot take over complex daily tasks. It does not run large-model inference itself, instead relying on the user's existing ChatGPT and Claude accounts, and uses idle computer time overnight to read and organize emails, files and chat logs locally into a personal knowledge base. According to the report, once that knowledge base exists, a simple instruction such as "finish this for me" lets it take over on-screen actions, while a double-tap shortcut can draft a reply in about two seconds using historical context. The project is built by CEO Jesai Tarun and CTO Aditya Vellanki, is based in San Francisco with a two-person team, has joined Y Combinator (F26 batch) and has raised more than $1M in pre-seed funding; the team argues that benchmark scores matter less than understanding the user.
DoNews · 🔥 0 - Meta, OpenAI and Google race into personal agents as YC's Garry Tan touts 'personal AGI'
Personal AI agents have become the hottest product category in AI over the past two months, according to an InfoQ roundup: Meta launched Muse, which it calls a "personal AI agent" with its own runtime and browser that can handle email, bookings, forms and long-running tasks while learning about the user, and later folded it into AI glasses, enterprise products and its developer platform. OpenAI's Dots gives each agent its own cloud computer that works toward user goals and connects to many external apps; xAI shipped Grokbot in August, and Google and Microsoft pushed their products and operating systems toward agents in October. In a Startup School 2026 talk, YC CEO Garry Tan argued that "personal AGI" will not arrive as a single event but as something diffuse — a terminal window, Markdown files and an agent running while you sleep. He described building his own system on roughly 220,000 Markdown pages, Skill files and the open-source GBrain framework, and claimed his coding output is about 400x his 2013 level, or 8x even under the harshest discounting.
InfoQ 中文 AI&大模型 · 🔥 0 - Vitalize raises $31M Series A led by Oak HC/FT for hospital staffing AI
Vitalize, an AI platform that helps hospitals manage labor and capacity in real time, raised a $31 million Series A led by Oak HC/FT, with Y Combinator and others participating, per Fierce Healthcare as relayed by Techmeme. The report describes hospital staffing and scheduling as unglamorous, complex and tedious work, and identifies Veeraj Shah as co-founder and CEO of the startup. The sources do not specify how the funds will be used, the company's valuation, or how many hospitals it serves.
Techmeme · 🔥 0 - Nous Research raises $90M at $1.5B valuation for its Hermes AI agent
Nous Research, the company behind the open-source Hermes agent framework, has raised $90 million at a $1.5 billion valuation in a round led by Robot Ventures, with investors including Nvidia, Microsoft's M12, Samsung and Y Combinator, according to a Cyzone report; cumulative funding now stands at $158 million. Nous says Hermes, released under the MIT license, has been cloned more than 24 million times, and it estimates that roughly 2.5% of global AI token usage comes from Hermes-related applications. Tencent's Lightweight Cloud team built LightVela on the Hermes agent with WeChat access, and Xiaomi's MiMo was an important model source for Hermes on OpenRouter; Nous reported about $36 million in annualized revenue as of mid-September 2026 and expects to top $100 million by year-end.
创业邦 科技 · 🔥 0 - SpaceX resells NVIDIA GPUs to Anthropic as Silicon Valley's compute crunch worsens
A QbitAI roundup drawing on The Wall Street Journal, Anthropic's blog and SemiAnalysis data says Silicon Valley is locked in a scramble for compute: SpaceX is acting as a reseller of NVIDIA GPUs, supplying capacity to Anthropic, and its first earnings report showed quarterly AI business revenue of $2.561 billion, up 247% year over year. The piece says Anthropic's May agreement with SpaceX added more than 300 MW of new capacity and 220,000 NVIDIA GPUs and raised Claude Code and API usage limits. It also reports that Noam Shazeer, hired back by Google for $2.7 billion, moved to OpenAI after compute was reallocated (citing foreign media), that H100 one-year rental prices rose from $1.70 per GPU-hour in October 2025 to $2.35 in March 2026 — nearly 40% — and that startups face multi-year contracts and prepayments of up to 30%.
量子位（原生 RSS） · 🔥 0 - Flai raises $27M Series A for automotive AI customer service
On October 8, Flai, which provides AI customer service for the auto industry, announced a $27 million Series A led by Base10 Partners, with Findlay Automotive Group, The Friedkin Group and Y Combinator among the participating investors. Flai offers auto dealers a voice-first, omnichannel reception platform that answers calls around the clock for pre-sales questions, test-drive bookings, service scheduling and recall notices, and integrates with DMS and CRM systems for automatic conversation logging and seamless handoff to human staff in complex cases. The funding is earmarked for product development and market expansion.
DoNews · 🔥 0 - Pentagon's Tradewinds uses 5-minute videos to speed AI 'kill chain' buys
Tradewinds, a procurement marketplace run by the Pentagon's Chief Digital and Artificial Intelligence Office (CDAO), lets companies submit product videos of no more than five minutes describing how their AI offerings address periodically updated "strategic focus areas"; a panel reviews submissions at least monthly, and accepted products gain "post-competitive" status that makes it easier to buy them via vehicles such as Other Transaction agreements. A DOD official says some awards have been made in under a week, and OpenAI, Anthropic and Google are all listed as participants. WIRED reports that 2026 topics seek AI that can "enhance force lethality" and "unleash" AI-enabled agents for joint targeting, including so-called kill chain execution, while earlier pitches on a "culture of responsible AI" and guardrails are no longer sought.
WIRED AI · 🔥 0 - Nous Research raises $90M at $1.5B valuation, launches Hermes for Businesses
Nous Research, maker of the open-source Hermes agent, raised a $90 million Series B led by Robot Ventures, with participation from Nvidia, Samsung, Union Square Ventures, Menlo Ventures, Y Combinator and 1789 Capital (The Next Web also lists Microsoft's M12). TechCrunch and The Wall Street Journal put the valuation at $1.5 billion, while Techmeme's headline says $1.2 billion; the company did not disclose a valuation itself. The money will fund "Hermes for Businesses," aimed at letting companies pick models by price and goal, keep AI usage private and own their company knowledge. Nous says Hermes has been cloned more than 24 million times since its February MIT-licensed release and drives roughly 2.5% of global AI token usage — its own estimate — while the Journal reported about $36 million in annualized revenue by mid-September and an expectation of passing $100 million by year-end.
Techmeme · 🔥 0
Experience and discussion from the community
Share my Y Combinator experienceAsk about Y Combinator
Nobody has shared their experience with Y Combinator yet.