Hello hello!
Welcome to another edition of Why Join. I keep thinking about this thing. The people who win never look like they're trying that hard. And the ones trying the hardest keep tripping over themselves. Weird how that works. Anyway lets jump in.
You know that app idea you’ve been sitting on? Anything lets you build it by literally just... describing it. Tell it what you want, and it spins up a full-stack app (mobile, web, or both) with a real backend, database, payments, the works. No code. Over 1M people are already building with it, and some of them have shipped actual App Store apps without writing a single line.
We got WHY JOIN readers 30,000 free credits to try it out.
💵 Raised - $17M seed at $100M post-money valuation, led by Andreessen Horowitz and The General Partnership
One-liner: AI workspace for financial modeling - like Cursor but for spreadsheets.
Financial modeling is painfully manual. Analysts spend hours building valuation models, DCFs, all that. And the thing is, unlike software engineering where ten different implementations are fine, in finance ten analysts modeling the same company should get basically the same workbook. That’s a huge tension with AI, because LLMs are inherently non-deterministic. You ask the same question ten times, you get ten different answers. Existing tools (like Shortcut AI) just bolted agents onto Excel, which... kind of limits what you can do. You're still stuck in Excel's world. The gap between "AI flexibility" and "auditability" has kept most financial teams from trusting AI tools.
Meridian built their own IDE-style workspace from scratch. That lets them integrate data sources natively and layer conventional tooling on top of the agentic stuff to keep outputs auditable. You can see exactly where every assumption comes from, how the logic flows, removing what they call the "doubt layer" around LLM outputs.
💵 Raised - $30M Series A, co-led by Chemistry and Headline
One-liner: AI autopilot for manufacturing procurement - sourcing suppliers, negotiating, tracking orders, handling payments, all the way through.
If you're a manufacturer buying raw materials from suppliers across, say, 15 countries, that whole process is just... chaos. It's emails, WeChat threads, phone calls, purchase orders, packing lists. People literally spend their entire day chasing suppliers and copy-pasting updates into ERPs. The existing AI procurement tools (Zip, Levelpath, SAP Ariba, etc.) mostly handle corporate purchasing, like buying software licenses or office supplies. Not the same thing. And the smaller players that do touch supply chain only really work for small brands and don't cover the full quote-to-payment flow.
Didero plugs into your existing ERP as an agentic layer and basically reads all that messy communication, the emails, messages, documents, then does the updates and tasks automatically. In Spencer's (CEO and co-founder) words, “the goal is to go from ‘I need a good’ to payment without having to lift a finger.”
Hiring: Senior Customer Success Manager, Growth Marketing Lead, Senior Product Manager - NYC/Remote
Anthropic raises $30B at a $380B valuation. Series G led by GIC and Coatue. Co-leads include D.E. Shaw Ventures, Founders Fund, ICONIQ, MGX. Plus basically every name you’ve heard of: Sequoia, Lightspeed, Fidelity, BlackRock, Blackstone, Goldman, JPMorgan, Jane Street, Temasek, you get the idea. The numbers are wild. $14B run-rate revenue, 10x annual growth for three consecutive years. 500+ customers spending $1M+ annually (that was a dozen two years ago). Eight of the Fortune 10 are Claude customers. Claude Code hit $2.5B run-rate on its own and has doubled since January. Anthropic says 4% of all public GitHub commits are now authored by Claude Code. First dollar of revenue was less than three years ago.
Source: Anthropic OpenAI drops Codex-Spark, 15x faster coding on Cerebras chips. GPT-5.3-Codex-Spark is a smaller, speed-first version of Codex built for real-time back-and-forth coding rather than the slow batch-style agent workflow. 15x faster code gen, 80% less roundtrip latency, 50% faster time-to-first-token. Runs on Cerebras’ Wafer Scale Engine 3, making it the first real output from their January partnership. The tradeoff: it’s less capable than full GPT-5.3-Codex on benchmarks and OpenAI says it doesn’t hit their cybersecurity capability threshold. So faster, but not as smart. Pro tier only ($200/mo) for now. OpenAI says they’re working toward blending both modes so you don’t have to choose. Not there yet.
Source: Cerebras Crypto flows to suspected human trafficking up 85% YoY. Chainalysis released part of its 2026 Crypto Crime Report and it’s grim. Hundreds of millions flowing to suspected trafficking ops, mostly tied to Southeast Asian scam compounds and Telegram-based laundering networks. Telegram “escort” services are highly professionalized, nearly half of transactions over $10K. CSAM operations shifted to subscription models under $100/month and are increasingly overlapping with extremist communities targeting minors. One positive: blockchain transparency actually gives law enforcement way better tracking than cash. German authorities shut down “KidFlix” and US arrests of CSAM buyers went up in 2025.
Source: Chainalysis TikTok rolls out Local Feed in the US. New location-based feed shows nearby restaurants, events, shopping, local news. Opt-in, 18+ only, location sharing off by default. Timing’s not great though. It showed up right after updated terms of service said TikTok would start collecting precise location data under the new US joint venture. TikTok is framing this around small business value (7.5M businesses on the platform, 28M jobs supported). Basically the same argument Meta makes when regulators come knocking. Already live in UK, France, Italy, Germany since December.
Two more xAI co-founders out the door. Tony Wu and Jimmy Ba resigned, which means Musk’s AI company has lost half its 12 co-founders in under three years. The FT says Ba left after internal friction over model performance pressure as xAI tries to catch up to OpenAI and Anthropic. Comes right after SpaceX said it would buy xAI to create a $1.25T combined company planning to IPO this year. Oh and at an all-hands, Musk floated building a factory on the moon to make satellites for AI. No timeline, no roadmap. Two former SpaceX execs told the NYT the moon was never really a priority there. Lot happening at xAI right now, not all of it reassuring.
Zhipu AI launches GLM-5. The Chinese AI company (also known as Z.ai) hit #1 on Artificial Analysis’s Intelligence Index for open-weights models, scoring 50. First time any open model has crossed that mark. 744B total parameters, 40B active, trained on 28.5T tokens, uses DeepSeek’s Sparse Attention, MIT licensed. Ranks third overall on agentic tasks behind Claude Opus 4.6 and GPT-5.2. Cut hallucination rate 56% by basically just saying “I don’t know” more often. Can run on Chinese-made chips from Huawei and others, notable since Washington blacklisted Zhipu last year. Stock jumped 29% in Hong Kong. Available via APIs around $1/$3.2 per million tokens. Part of a bigger wave of Chinese model drops ahead of Lunar New Year.
Mistral investing €1.2B in Swedish data centers. CEO Arthur Mensch announced Mistral’s first data centres outside France at the Techarena conference in Sweden. Partnering with EcoDataCenter, chose Sweden for clean energy. Mistral is valued around €11.7B, claims 20x growth over the past year, targeting €1B+ revenue in 2026. Mensch pushed hard on Europe needing to act as one market to compete with the US.
Handshake: Software Engineer - Intern - San Francisco
Hook: Sales Development Representative (Graduate) - London
Thera: Sales Development Representative (SDR), Founding Engineer - NYC
The Compression Company: ML Compression Engineer, Deployment Engineer, Forward-Deployed Engineer, Earth Observation - US
Reflow: Customer Success Manager, GTM Manager, Product Manager - Remote - US; Software Engineer, AI Engineer - Remote - Anywhere in the world
Seamflow: Founder’s Associate, Founder’s Associate - Intern - London
Stanhope AI: Chief of Staff, AI Developer - Robotics & Embodied Active Inference Systems - London
Take2 AI: SDR, Account Executive, Head of Growth, Senior Software Engineer, Fullstack - New York
Anterior: Member of Technical Staff - New York
Ando: Fullstack AI Engineer, Chief of Staff - San Francisco
Ever: Product Designer, Software Engineer, Head of Brand & Marketing, Head of People - San Francisco
If you haven't yet, please subscribe, like, leave a comment, and share it! It helps us continue to bring you the most interesting news + nuance in tech and business every week.
See you Monday, Ryan
Sponsorships We are now accepting sponsors for Q1 '26. If you are interested in reaching my audience of founders, investors, and tech executives, send me an email at chief@whyjoin.xyz.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.