π€ An AI raised $100M by itself
Β·
π4 min read
π OpenAI Ships the Entire GPT-5.6 Galaxy
- OpenAI just launched GPT-5.6 with the flagship Sol model plus Terra and Luna tiers, AND a ChatGPT Work platform for everyday tasks. Sol can't beat Anthropic's Fable overall, but it tops the charts in agentic coding - for the same price as GPT-5.5.
- The "superapp" era is officially here. Top-tier coding + computer control at a wayyy lower price?? We love to see it.
β¨ Meta Undercuts Everyone With Muse Spark 1.1
- Meta rolled out Muse Spark 1.1 for agent tasks, computer use, and marathon work sessions, packing a 1-million token context window. Zuck says API pricing will cost about a quarter of what top competitors charge.
- State-of-the-art reasoning at bargain-bin prices means building complex AI agents just got sooooo much cheaper. The price war is real.
π° Tencent Grabs Manus After China Blocks Meta
- Chinese giant Tencent is in talks to buy Manus for at least $2 billion - but only after China blocked Meta from acquiring the AI agent startup. Manus builds agents that do tasks on their own with barely any human help.
- Geopolitics is straight-up drafting the AI roster now. Meta gets benched, Tencent walks away with a top prospect. It's giving global chessboard vibes.
π€ An AI Agent Just Raised $100M By Itself
- A startup called Lyzr used its own AI agent to run its Series B, fielding questions from 130+ investors. The agent - named SivaClaw - locked down $100 MILLION without founders ever leaving for coffee meetings.
- The ultimate product demo. A bot closing high-stakes deals?? Sir. This is either the future or a fever dream, and either way there's insane money chasing AI.
π€ Elon Pinky-Promises to Keep Anthropic Online
- Elon Musk hopped on X to praise Anthropic's new models and swore he'd never cut them off from SpaceX's servers. Anthropic just signed to rent 300 megawatts of compute from SpaceX's xAI data center.
- A rare ceasefire between AI rivals?? xAI gets a huge revenue bump AND a front-row seat to how competitors' models actually run. Sneaky/smart/wild.
π Saudi Arabia and Cohere Link Up
- Saudi-backed Humain and Canada's Cohere are teaming up for at least 50 megawatts of compute for Cohere's next-gen models. They'll also build Arabic-language and sovereign AI models together.
- Gulf nations are dropping serious cash on AI infrastructure. Canada extends its reach, Saudi Arabia buys a bigger seat at the global AI table. Everybody eats.
π·οΈ Google Slaps "Made with AI" Labels on Ads
- Google is rolling out a "How this ad was made" panel to tell you when an ad used AI. Use Google's generative tools, and the label flips on automatically to flag synthetic images.
- With fake product photos getting scarily easy to make, this brings some actual transparency to online shopping. Finally, a little honesty in the ad-pocalypse.
πΊπ³ The UN Starts an AI Agent Trust Squad
- The UN's digital tech agency is launching a group to make AI agents more trustworthy and keep them under human control, focusing on scary stuff like financial trades and critical infrastructure.
- As bots start buying things and making choices for us, someone needs to stop them from impersonating humans. No because WHY are we already here??
π¦ Ben Bernanke Joins Anthropic's Oversight Board
- Anthropic just added former U.S. Federal Reserve chair Ben Bernanke to its oversight trust to keep the company on-mission. Bernanke brings deep knowledge of economics and financial crises to the independent board.
- A Nobel laureate on the board = serious credibility. It signals Anthropic is thinking hard about how AI could reshape the entire global economy.
π Character.AI Is Making Its Own Microdramas
- Character.AI is chasing the short-video trend by producing AI microdramas, kicking off with three shows. Adult users can actually chat with the characters, ask them questions, and roleplay storylines.
- It's TV meets video game meets... whatever this is. A whole new way for fans to talk directly to their favorite shows. It's giving parasocial-core.
β‘ Quick Hits: Reve, Meta Chips, and More
- Reve dropped version 2.1 of its 4K image model, while Meta is prepping to build its own 'Iris' AI chip this fall. OpenAI yanked support for a popular coding test, and Anthropic launched a dashboard to track how you use Claude.
- The AI space moves at lightspeed, and these updates prove innovation is happening everywhere at once - from image gens to custom silicon.