TLDR 2026-07-06

NVIDIA TAPS AI CLOUD PROVIDERS TO EXPAND COMPUTE ACCESS FOR STARTUPS

Nvidia has started a partnership program that offers firms access to critical computing infrastructure through the company's network of AI cloud service providers. Members of the program will share both product and cloud revenue with Nvidia. AI firms are increasingly entering into revenue and equity-sharing agreements with chip makers to circumvent liquidity issues. Nvidia plans to raise at least $20 billion in debt for general corporate purposes.

AMAZON HAS DEPLOYED ENOUGH SATELLITES TO LAUNCH LEO SERVICE LATER THIS YEAR

Amazon Leo now has enough satellites in orbit to begin serving customers. The constellation now has more than 390 satellites. Amazon began offering an enterprise preview for select businesses last year. Its commercial service will likely initially be limited to users in certain geographic regions.

THE QUEST TO MAKE HUMANOID ROBOTS SAFE ENOUGH FOR HUMANS

One of the biggest challenges for robot makers is ensuring that they don't hurt humans. The machines are getting bigger and heavier, which increases their potential for damage. The International Organization for Standardization is expected to publish a standard for humanoid robot safety by mid-2028, but robot makers are already coming up with their own solutions. This article looks at some of those solutions.

THE AI SUPERFORECASTERS ARE HERE

An AI superforecaster is an AI modified to be good at forecasting. This usually involves a scaffold that handholds the agent through a long research process. There are now several companies that claim to have AI superforecasters with incredible performance and results. They appear to be making crazy profits on prediction markets and beating the stock market by some comfortable amount.

pay attention. Learn more about sponsorship opportunities.

AGENTIC AUTONOMY LEVELS

Low autonomy limits risk and increases reversibility. Higher autonomy is better for explicit activities, as well as for fleets of parallel agents refactoring massive codebases. The edge of the frontier is the manager agent that delegates to helpers while continuously verifying their output, and only returning with the decisions that must be made by a human. This setup can be used to run hundreds or thousands of agents.

A FIELD GUIDE TO FABLE: FINDING YOUR UNKNOWNS

Users need to spend more time defining unknowns to prevent long-horizon tasks from coming back wrong. Claude can be used to help define these unknowns through explainers, brainstorms, interviews, prototypes, and references. Doing this is a cheap way to fix problems before they get expensive. This post shows readers how to use Claude to help find their unknowns.

THREADS, META'S ‘TWITTER KILLER,' FINDS ITS PEOPLE

Threads now has 500 million monthly users. The app has become especially popular in Asia. Threads is now targeting one billion users. The app started showing ads in January - reaching its user goal could lead to revenue of at least $30 billion a year.

AI HAS TORCHED THE MARKET FOR JUNIOR PROGRAMMERS

Get $1 off your domain name at Porkbun

BETTER MODELS: WORSE TOOLS

If the newest models get better at solving the task while getting worse at faithfully emitting an alternative tool schema, then the harness needs stronger guarantees somewhere.

WORDPRESS DROPS TO 41.5% MARKET SHARE, BUT THREE DATASETS TELL DIFFERENT STORIES ABOUT WHAT'S GOING ON

WordPress now powers 41.5% of the web, down from a peak of 43.6% in early 2025.

AGENTIC LOOPS

A look at various strategies to run agentic loops.

APPLE PLANS FIVE NEW IPHONES THROUGH 2027, EYES CHINESE-MADE CHIPS AMID FOLDABLE PUSH

Apple has instructed suppliers to prepare to produce about 10 million foldable iPhones this year.

DOES A URL IN A PROMPT STEER AN LLM'S OUTPUT TOWARD ITS CONTENT?

When a URL and its content make it into the training data, the bare URL works as a key into the weights

SOME NEW AGENTIC PATTERNS

A look at how Prime Radiant uses agents.