Daily Edition · Page 1

THE DAILY INDEX

Technology, AI, software, security, crypto and business

Across today’s newsletters

Today’s Index

APPLE TO LAUNCH IOS 26, MACOS 26 IN MAJOR REBRAND TIED TO SOFTWARE REDESIGNS

The next Apple operating systems will be identified by year rather than version number. The change is aimed at bringing consistency to the company's branding and making things less confusing to customers and developers. The next operating systems will be known as iOS 26, iPadOS 26, macOS 26, watchOS 26, tvOS 26, and visionOS 26. The shift will be announced at Apple's Worldwide Developers Conference on June 9.

NYC MAYOR CALLS FOR END OF NY BITLICENSE, PROPOSES 'BITBOND'

NYC Mayor Eric Adams called for the abolition of the New York Department of Financial Services BitLicense to entice crypto firms back to the city in a Bitcoin 2025 keynote. He proposed a “BitBond” municipal bond backed by bitcoin to raise capital with tax advantages, which would allocate 90% of proceeds to government funding and 10% to Bitcoin purchases. It would pay holders 1% annual interest for 10 years plus 100% of Bitcoin's upside up to a 4.5% compounded return, with residual gains seeding a city Bitcoin reserve.

RESEARCHERS SAY NEW CONTACT LENSES LET YOU SEE IN THE DARK

Scientists have created contact lenses that allow people to see in the dark with the help of infrared light. The invention could allow colorblind people to see colors again. The contacts don't require an external power source and wearers can perceive multiple infrared wavelengths even with their eyes closed. They aren't yet able to capture fine details, so the scientists are working on a version with more precise spatial resolution and higher sensitivity.

A REACTION TO ALZHEIMER'S FRAUD

The beta-amyloid protein has been a prime suspect in the pathology of Alzheimer's for many decades. However, there was a period where some papers made a form of amyloid appear to be able to induce disease-like symptoms all by itself in rodent models. Those works turned out not to be reproducible, and there is strong evidence of deliberately faked work in those papers. While those particular papers are fake, they don't discount the decades of legitimate research on the topic.

ONE OF THE REASONS I'M FINDING LLMS SO MUCH MORE USEFUL FOR CODING THAN A LOT OF PEOPLE

All of Simon Willison's code has automated tests. This massively de-risks his use of large language models. He uses AI to write code and then proves that code works with tests, then uses AI to refactor the code until it looks good, keeping the tests green the whole time. Programmers who find the technology to be more of a hindrance than a help in their work may need to check the health of their test suites.

TLDR IndexPage 1