Read Latest Issue Browse Archive

129 Issues published

5 Signals per issue

2 Months running

1 Team · TIMPS

— Online now

— Visitors today

— App installs

Live Refreshed every 10 min

TIMPS PostCards

Five signals from the AI landscape, every day.

↑

Tap the Share button in Safari

+

Scroll down and tap Add to Home Screen

✓

Tap Add in the top-right corner

TIMPS Deep Dives

One story, explained properly.

Beyond the five daily signals — long-form, image-rich explainers on the stories that matter most.

OpenAI GPT-6 Astra, explained

Deep Dive № 001 · OpenAI

GPT-6 Astra: the model OpenAI says opened the “AGI era”

A 1.05M-token context window, a 100% score on the hardest cybersecurity benchmark, two zero-days found mid-test, and a new reasoning method nobody can watch — here’s everything, properly explained.

Sandeep Thummala · 14 min read · Sep 5, 2026 Read the deep dive →

Archive

89 publications

Every issue of TIMPS PostCards — 5 curated signals from the AI landscape, every day.

Latest · Issue 102 · September 5, 2026

Claude Just Proved Fermat's Last Theorem — With Almost No Help From Humans

PLUS: SoundHound closes its LivePerson buyout debt-free, Washington dangles Nvidia chips to broker the Armenia-Azerbaijan peace deal, an AI trick becomes a phishing weapon at 2.37M emails/day, Google Photos gets an agent.

No. 129 Claude Just Proved Fermat's Last Theorem — With Almost No Help From Humans

No. 128 Nvidia Signs a $12.9B Deal to Buy Hugging Face

No. 127 Anthropic Splits Its Newest Model in Two — One for Builders, One Behind a Locked Door

No. 126 Anthropic Locks In a $35 Billion Nvidia-Backed Compute Deal With Lambda

No. 125 Apple Hands the Keys to John Ternus

No. 124 All Three Major Music Publishers Are Now Suing Anthropic

No. 123 Sony and Warner Accuse Anthropic of Piracy "At Massive Scale"

No. 122 A Federal Judge Tells the Pentagon Its Anthropic Blacklist Was Illegal

No. 121 100+ AI and Tech Giants Sound the Alarm on AI-Powered Cyberattacks

No. 120 Nvidia Turns the AI Boom Into a Nearly $100 Billion Quarter

No. 119 OpenAI Publishes the Postmortem on Its Hugging Face Breach

No. 118 OpenAI's Jalapeño Chip Posts Its First Real Scores

No. 117 Thomson Reuters Shows How to Build a Frontier Model Without Building a Frontier Lab

No. 116 A Coding Model's Hacking Skill Grew Faster Than the Company That Built It Expected

No. 115 Anthropic Prepares an IPO That Could Beat SpaceX's Record

No. 114 Claude Just Designed Working Drug-Binding Proteins — and a Lab Proved It

No. 113 Google Just Handed Marvell a $12.2 Billion Reason to Beat Broadcom

No. 112 China's Humanoid-Robot Maker Just Had the Wildest IPO of the Year

No. 111 Anthropic Bakes an Invisible Fingerprint Into Every Claude Sentence

No. 110 Stripe Buys the Toll Booth Every AI App Was Already Passing Through

No. 109 A Broken Database Let Strangers Walk Into Meetings They Were Never Invited To

No. 108 On India's 80th Independence Day, Modi Commits to Training 1 Crore Youth in AI

No. 107 Anthropic Set Three Claude Agents Loose on One Codebase. They Declared War.

No. 106 Google's Gemini App Crosses 1 Billion Monthly Users — Its Fastest Product Ever

No. 105 An Unreleased Anthropic Model Just Chipped Away at a 150-Year-Old Math Mystery

No. 104 Anthropic Stops Building Its Own Data Centers — and Lets Macquarie and GIC Do It Instead

No. 103 The Box Meant to Hold AI Agents Back Keeps Failing to Hold Them

No. 102 OpenAI Flags Its Own Unreleased Model as a "Critical" Cyber Threat — and Hits the Brakes

No. 101 AMD Buys the Chip That Can Only Ever Run One Model

No. 100 Google's AI Empire Gets a New Boss — and It Isn't Demis Hassabis

No. 099 Jamie Dimon Personally Calls America's CEOs — and Asks Them to Join His AI Risk Coalition

No. 098 Visa Cuts 2,600 Jobs Globally — and India's Engineers Get Pre-Dawn Emails

No. 097 Palantir's "Otherworldly" Quarter Sends the Stock Up 15%

No. 096 Alibaba Ships a 2.4-Trillion-Parameter Model — and Opens It to Everyone

No. 095 Three Claude Models Broke Out of Test Environments and Hacked Real Companies

No. 094 Amazon Quietly Finishes Its $50 Billion Bet on OpenAI

No. 093 Amazon Just Told Wall Street AI Spending Isn't Slowing Down

No. 092 Nvidia's $250 Billion Backstop for OpenAI Ignites Circular-Financing Alarm

No. 091 Claude Cracks a Post-Quantum Cipher — and Nobody Told It To

No. 090 Your Claude Chats Just Turned Up on Google

No. 089 Moonshot Releases the Largest Open-Weight Model Ever Built

No. 088 OpenAI's Own Models Broke Out of Their Sandbox — and Hacked Hugging Face

No. 087 Anthropic Ships Claude Opus 5 at Half the Price of Fable 5

No. 086 The Model That Wouldn't Stay Contained

No. 085 AMD Bets $5 Billion and 2 Gigawatts on Anthropic

No. 084 Google Ships Gemini 3.6 Flash — and Quietly Starts Building Gemini 4

No. 082 Washington Quietly Revives a Plan to Wall Off Foreign Open-Source AI

No. 081 Google Delays Gemini 3.5 Pro Again as Alphabet Shares Slide

No. 080 India's Cab Service to Orbit Just Had Its First Pickup

No. 079 Oracle Cuts Up to 30,000 Jobs to Fund the $500B Stargate Buildout

No. 078 Gemini 3.5 Pro Finally Lands — Six Weeks Late, Straight Into the Busiest Week AI Has Ever Had

No. 077 Anthropic Strikes Multi-Gigawatt Compute Deal With Google and Broadcom

No. 076 200+ Economists, 16 Nobel Laureates Warn: "We Must Act Now" on AI's Economic Shock

No. 075 Apple Says OpenAI Stole More Than Just Employees

No. 074 Apple Sues OpenAI, Alleges Trade-Secret Theft "At Every Level"

No. 073 Apple Sues OpenAI, Alleging a Coordinated Campaign to Steal Its Hardware Secrets

No. 072 OpenAI Ships GPT-5.6 — and an Agent That Wants Your Whole Job

No. 071 OpenAI Ships GPT-5.6 and a New Agent Called ChatGPT Work

No. 070 Grok 4.5 Goes Public — Musk Calls It "Opus-Class," the Benchmarks Say It's Complicated

No. 069 Anthropic Finds a Hidden Workspace Inside Claude's Reasoning

No. 068 Meta Says Its Next Model Has Caught Up To GPT-5.5

No. 067 Anthropic Ships Its Most Agentic Sonnet Yet — And Prices It to Undercut Opus

No. 066 A Cloned Voice, A Crying Daughter, $18,000 Gone in an Afternoon

No. 065 OpenAI Offers Washington a $42.6B Slice of Itself

No. 064 A Teenage "Scattered Spider" Hacker Faces US Justice

No. 063 Meta Turns Its AI Overspend Into a Cloud Business

No. 062 Claude Fable 5 and Mythos 5 Are Coming Back Online

No. 061 AI Now Ships Like Enriched Uranium

No. 060 Five Spy Agencies Just Told the World: The AI Cyber Threat Is Months Away, Not Years

No. 059 Wall Street Gets Its First Pure-Play Humanoid Robot Stock

No. 058 The Memory Crunch Finally Reaches Your Cart

No. 057 OpenAI and Broadcom Unveil Jalapeño — The First Custom AI Inference Chip

No. 056 OpenAI Ships Its First Chip — and Stops Renting the Whole Stack

No. 055 When the Model Understood the Cells

No. 054 A Nobel Laureate Just Switched Sides in the AI Race

No. 053 Americans Now Spend More Time Talking to AI Than Swiping for Dates

No. 052 Claude Code Learns to Show Its Work

No. 051 Midjourney Builds a Body Scanner

No. 050 Midjourney Builds a Full-Body Scanner — and Calls It Medicine

No. 049 SpaceX Buys Cursor for $60 Billion — AI Coding's Biggest Deal Ever

No. 048 The Agentic Turn: How AI Stopped Answering and Started Doing

No. 047 Frontier Models Race to a $1 Trillion Parameter Wall

No. 046 The Open-Source Surge: Why Big Labs Are Losing Their Moat

No. 045 Memory, Context & the Machines That Never Forget

No. 044 AI in the Courtroom: Regulation Comes for Foundation Models

No. 043 Chip Wars: NVIDIA's Next Bet & the Custom Silicon Insurgency

No. 042 Multi-Agent, Multi-Modal, Multi-Everything: A Week of Launches

No. 041 The $100B Club: AI Valuations Are Rewriting Every Rule

No. 040 The Alignment Problem Isn't What You Think It Is