Grow your business with AI tools and hacks?
Earn more, work less with AI.
Stay up-to-date with the latest AI tools, hacks, and prompts delivered straight to your inbox.
Want to sponsor a newsletter? Find out more.Plus: OpenAI's $1.25 trillion valuation, Anthropic's IPO countdown, and the open-source tool that catches model swaps
Sponsored by
Create scroll-stopping AI content with Higgsfield. Generate images, videos, voices, and cinematic visuals using powerful AI tools in one creative suite.
Try Higgsfield today and unlock your next big idea.Hey there,
This week AI officially left the planet. Google powered up its first Project Suncatcher satellite, sent a prompt to orbit, and got back a haiku. Meanwhile back on Earth, OpenAI's valuation hit $1.25 trillion, Anthropic inched toward its IPO, and a report dropped suggesting Anthropic's models are now improving twice as fast as a year ago — largely because AI is helping build the AI.

On this day in 2025, the AI industry was in the middle of its most compressed release window ever — sitting between OpenAI's then-record $38 billion AWS compute deal in early November and the Gemini 3 and GPT-5.1 launches that landed just days later. One year on, the frontier labs have swapped "bigger model" announcements for orbital TPUs and IPO filings.
🛰️ Google's first space data center satellite goes online — and writes poetry
Google's Project Suncatcher prototype, launched aboard a SpaceX rocket last month, completed the first AI response ever generated by a data-center-class TPU in orbit. Engineers sent Gemma 4 a prompt during a ground-station pass over Svalbard and got back: "Blue marble below / Data streams in silent grace / Starlight feeds my core." Sundar Pichai called it a "hello world" moment. Next milestone: two laser-linked satellites in 2027, sharing a single AI workload on solar power that's up to 8x more productive than on Earth. The catch — researchers estimate Starship would need ~1,800 flights to make launch costs work in the 2030s.
💰 OpenAI valued at $1.25 trillion as Anthropic nears its IPO
OpenAI closed new investment at a $1.25T valuation, with reports suggesting it's still chasing up to $20 billion total. Notably, it had originally hoped to raise as much as $30 billion at $1.4 trillion — before tech stocks slid and midterm results showed how sharply public opinion has turned against AI data center projects. Rival Anthropic filed its S-1 last week, with analysts pegging Nov. 16–19 as the likely debut window.
🧰 OpenAI is acquiring Pierre, the company behind Codex's code review tools
If you've browsed a file or reviewed a diff in the Codex app, you've already used Pierre's Trees and Diffs libraries. OpenAI is buying the company behind Code Storage to rebuild Git workflows around coding agents — following its reveal of Pando, an in-house source control backend built for engineers and agents working in parallel. Pierre's open-source projects stay open source, and OpenAI says it will keep contributing.
⚡ New METR report on Anthropic automating AI development is alarming researchers
Anthropic's models are improving roughly twice as fast as a year ago, with no observable bottlenecks or diminishing returns — and much of that acceleration traces back to internal models the public has never seen. METR identified three unreleased models (codenamed Model 2, 3, and 4) that finished training between July and October. Most of the speedup still comes from AI assisting human researchers, but METR warns there are "some signs that fully autonomous contributions will play a larger role soon."
🔐 anthropics/FlashProve: Anthropic open-sourced a research prototype that uses zkSNARKs to cryptographically prove which model generated a given response — so a provider can't quietly route your request to something cheaper. An offline step publishes a short commitment to the model weights, and anyone can then verify a proof on an ordinary CPU in seconds without ever seeing those weights. Ships with a GPU prover, a C++ verifier, a second verifier formally proved sound in Lean 4, and a vLLM middleware integration. Runs Llama 4 Maverick at full size on a single H100 at ~110 tokens/sec, though proving costs roughly 70–95x the compute of normal inference and context is capped at 1,024 tokens. Apache 2.0, explicitly unmaintained — Anthropic says it published this so others would "take it and improve it."
⚙️ Kinference: Buoyant (the team behind Linkerd) shipped an open-source inference control plane for vLLM on Kubernetes, now generally available under Apache 2.0. You point clients at an HTTP endpoint, send OpenAI- or Anthropic-compatible traffic, and get token rate limiting, per-tenant usage and cost tracking, intelligent load balancing, and declarative inference policies via Kubernetes CRDs — no client-side integration required. Buoyant's own benchmarks claim a 95% reduction in median time-to-first-token versus standard Kubernetes service load balancing, which is a vendor number worth verifying on your own cluster.
Coming in Hot
Find and compare AI tools for every ecommerce workflow.
AI text humanizer: rewrites AI writing so it reads naturally. 300 words free.
Free, open-source video editor for everyone
AI video platform for workplace training teams
Receipt autopilot for freelancers and small businesses
Pro video editing, color, VFX & audio in one app
SMS-native runtime for building AI agents
AI-powered tools for writing, PDF editing, visual creation, and smarter digital workflows.
One API for top image and video generation models
AI social media content generator for marketers
Submit your AI tool to the #1 AI tools marketplace.

What it does: GPTHumanizer is an AI text humanizer that rewrites AI-generated and existing drafts into more natural-sounding prose. It adjusts word choice, sentence structure, paragraph flow, and tone while preserving numbers, dates, links, and email addresses. Users can compare the rewritten version side by side with their original text before making changes.
Why It's Worth Your Time
Pricing: GPTHumanizer offers a free one-time allowance of 300 words after signing up, with no payment method required. Paid plans include Basic at $9.90/month for 15,000 words, Pro at $18/month for 50,000 words, and Ultra at $36/month for unlimited words. Annual billing is also available. All plans include access to the Humanizer, long-text workspace, and Essay Writer, with a shared word balance.
Why it matters: GPTHumanizer helps students, bloggers, content creators, and marketing teams refine existing drafts without rewriting everything from scratch. Its side-by-side editing and preservation of key details make it easier to review changes and retain the original meaning. However, it does not guarantee that content will pass AI detectors, so users should treat its output as an editing aid rather than an AI-detection bypass.

What it does: EcomAgentTools is an ecommerce AI tools directory that helps online store owners discover, compare, and choose AI tools, agents, reusable skills, and prompts for different business workflows.
Why It's Worth Your Time
Pricing: EcomAgentTools is free to browse and use. However, individual AI tools and third-party services listed in the directory may require paid subscriptions, usage fees, or additional integration costs.
Why it matters: EcomAgentTools helps ecommerce entrepreneurs, online retailers, Shopify sellers, and marketing teams find AI solutions for everyday business tasks. By organising tools around practical ecommerce workflows, it makes it easier to compare options, discover useful resources, and choose solutions that can save time, reduce manual work, and improve store operations.
🐶 Meta opened Muse Voice to all US users, letting anyone call or video chat with their animated AI avatar on mobile, web, and Mac — powered by new Muse Realtime Avatar and Realtime Voice models. Glasses and the Tamagotchi-like "Muse Charm" wearable are next, before year-end.
📱 Nothing's first "AI-native product" ships in December, and CEO Carl Pei is being deliberately cryptic: no new screen, couldn't have been an app, and it's a "productivity tool." He also took a swipe at the pendant-pin-glasses gold rush, writing that while he admires the boldness, "I read this moment differently."
🛻 Serve Robotics wants to be the AWS for robots, launching a Physical AI Services line that sells its connectivity stack, an edge-case data curator, and a real-world foundation-model data catalog built from its own sidewalk fleet. Bedrock Robotics is already a customer; T-Mobile is helping push compute closer to the robots.