Search the archive

EVERYTHING APPLE PLANS TO SHOW AT TUESDAY'S ‘AWE DROPPING' IPHONE 17 EVENT

Apple plans to announce its next-generation iPhone lineup, new smartwatches, and other peripherals at a product launch event tomorrow. There will be four new iPhone models, including what will likely be called the iPhone 17 Air, a model approximately one-third thinner than the iPhone 16 Pro. It will have the same A19 processor as the iPhone 17, but can only support eSIMs due to its thin frame. More about what to expect from tomorrow's event is available in the article.

HOPE FOR DIABETES: CRISPR-EDITED CELLS PUMP OUT INSULIN IN A PERSON – AND EVADE IMMUNE DETECTION

Sana Biotechnology, a firm in Seattle, Washington, has successfully implanted CRISPR-edited pancreas cells into a person with type 1 diabetes without the need for immune-dampening drugs. The cells, collected from a deceased donor, were able to produce insulin for months. Currently, the only way for someone with type 1 diabetes to avoid dependence on injected insulin is through the transplantation of cadaveric islet cells, but this requires lifelong immune-suppressing drug therapy. Sana's solution bypasses the need for those drugs entirely.

ELON MUSK: NEURALINK COULD RESTORE PARTIAL VISION TO THE BLIND NEXT YEAR

Neuralink could attempt to restore limited sight to visually impaired patients as early as 2026. Blindsight is a brain-computer interface designed to restore vision. It targets the brain's visual cortex to generate visual perception, which means that vision could be restored even for people who were born blind. Blindsight has received a 'breakthrough device' designation from the US FDA.

HOW TO BUILD WITH NANO BANANA: COMPLETE DEVELOPER TUTORIAL

Google's Nano Banana model introduces state-of-the-art capabilities for creating and manipulating images. This guide provides a walkthrough on how to integrate the model into applications using the Gemini Developer API. It covers how to set up environments, generate and edit images, and apply advanced techniques. It also discusses how to use Nano Banana in AI Studio, best practices, effective prompting, and more.

ENGINEERING EXCELLENCE STARTS ON EDGE

The best engineering teams develop their own tools while running production code on edge, where progress is made. When programmers are able to directly influence the tools they're working with, they're likely to do so, which makes them learn more. However, this requires being able to immediately use the improvements or bug fixes they helped devise - it doesn't work if they have to sit and wait for the next release. Bug-free frameworks and libraries take work, but the reward is increased engineering excellence.

OPENAI BACKS AI-MADE ANIMATED FEATURE FILM

Chad Nelson, a creative specialist at OpenAI, is making a feature-length movie using AI. Expected to be released in theaters globally next year, Critterz is a story about forest creatures who go on an adventure after their village is disrupted by a stranger. Nelson has teamed up with production companies in London and Los Angeles to debut the film at the Cannes Film Festival in May. The team is attempting to make the movie in about nine months instead of the three years it would typically take. It has a budget of less than $30 million, far less than what animated films typically cost.

NO, AI IS NOT A BUBBLE

Bubbles are false beliefs about reality that everyone eventually figures out were false. The dot-com bubble that popped was the idea that businesses would become instantly rich just by being on the internet. People no longer believe that just adding AI to products will make them instant millionaires. The idea that modern AI will significantly transform business and the global economy, leading to massive unemployment for knowledge workers, is not a bubble.

the report

BEING GOOD ISN'T ENOUGH

The most important thing is agency - high-agency people make things happen, while low-agency people wait.

TESLA NOW SELLING SUPERCHARGERS DIRECTLY TO BUSINESSES

Business owners are now able to purchase and install Superchargers on their properties, with Tesla fully managing operations.

GOOGLE AI MODE MAY BECOME THE DEFAULT GOOGLE SEARCH EXPERIENCE “SOON.”

AI Mode becoming the default can mean big changes to the future of SEO.

DISTRIBUTING YOUR OWN SCRIPTS VIA HOMEBREW

This post walks through how to create a tap and a working formula for publishing to Homebrew - once the initial setup is complete, publishing additional tools becomes almost trivial.

LEAKED ANIMATION SHOWS HOW SAMSUNG'S TRIFOLD PHONE WILL CLOSE

The screens fold out and close in a specific order.

WHY WE TRUST STRANGERS' OPEN SOURCE MORE THAN OUR COLLEAGUES'

Employees often maintain their projects because they love them, so it's worth giving their projects the same fair evaluation given to any other open source.

⏰ EARLY BIRD ENDS SEPTEMBER 17. REGISTER NOW TO SAVE $400 →

ANTHROPIC PAYS $1.5 BILLION TO SETTLE AUTHOR COPYRIGHT LAWSUIT

A federal judge ruled in June that it was legal for Anthropic to train on copyrighted materials but that it was illegal to acquire the materials from pirate sites like LibGen. Anthropic has offered to settle the class action suit from authors by offering $3,000 per work.

ASML BECOMES MISTRAL AI'S TOP SHAREHOLDER AFTER LEADING LATEST FUNDING ROUND

ASML is committing 1.3 billion euros to Mistral's 1.7 billion euro fundraise, making it set to become Mistral's top shareholder. It is expected to get a board seat at Mistral. The funding round will make Mistral the most valuable AU company in Europe. The cash could help Mistral make Europe less reliant on US and Chinese AI models.

OPENAI EXPECTS BUSINESS TO BURN $115 BILLION THROUGH 2029

OpenAI has raised its projected cash burn through 2029 to $115 billion, $80 billion higher than the company previously expected. It will burn more than $8 billion this year, $1.5 billion more than it projected earlier this year. The company plans to develop its own data center server chips and facilities to power its technology and control its soaring costs. Its first chip, developed in partnership with Broadcom, will only be used internally and won't be made available to customers.

WHY LANGUAGE MODELS HALLUCINATE

Hallucinations persist because current evaluation methods reward wild guessing that might occasionally be correct over admitting a lack of knowledge. The fix is simple: calibrate rewards during training based on certainty, penalize confidently wrong errors more, and give partial credit when models hedge or express doubt.

DON'T BUILD AN AI SAFETY MOVEMENT

Building a popular AI safety movement risks undermining existing organic public support by making it appear coordinated rather than grassroots. What has worked so far is authentic opposition to specific policies and elite-focused expert advocacy. The author argues that broad movements inevitably get captured by adjacent political causes like labor or environmental issues, diluting safety-specific goals while exposing credible organizations to guilt-by-association with more extreme protesters.

COMPUTE SCALING WILL SLOW DOWN DUE TO INCREASING LEAD TIMES

The era of rapid compute scaling that drove AI progress since 2020 is over. GPUs from large cloud providers are almost entirely saturated, and frontier labs are now racing to bring new data centers online and reserve as much chip production as they can. Scaling, and therefore investment, is hampered by physical reality and the years required to build more infrastructure, meaning incremental scaling is the only option ahead, despite headlines of massive upfront funding rounds.

Agentic AI. Get the eBook

THE RACE TO BUILD A DISTRIBUTED GPU RUNTIME

NVIDIA and AMD have been racing to solve data movement at the cluster scale to create a software moat. Voltron Data's Theseus is the first distributed runtime to put movement at the center of scheduling, memory, and I/O, and then prove the payoff with benchmarking results. It is open, composable, and already running on CUDA and ROCm. Theseus runs across both Nvidia and AMD ecosystems for datacenter-scale analytics and AI.

CLAUDE CODE FRAMEWORK WARS

Dozens of open source projects are currently testing different recipes for how to work with AI productively. These efforts show that AI works best when it is given structure. The frameworks are converging on a future where AI becomes a set of teammates to be managed. The more structure you give, the more you get back.

QWEN3-MAX-PREVIEW ANNOUNCED

Alibaba has introduced Qwen3-Max-Preview, a 1 trillion-parameter model now accessible via Qwen Chat and Alibaba Cloud. Early benchmarks highlight improvements in instruction following, dialogue, and agentic task performance.

THE AGENTIC SYSTEMS SERIES (BOOK)

The Agentic Systems Series is a complete guide for building AI coding assistants that actually work in production. It covers everything from fundamental concepts to implementing enterprise-ready collaborative systems. The series provides deep analysis of real production systems, including Amp, Claude Code, and open-source implementations like anon-kode. It is assumed readers are familiar with system design concepts and have a basic understanding of AI/LLM integration and experience with either TypeScript/Node.js or similar backend technologies.

LLM TRAFFIC: WHAT'S ACTUALLY HAPPENING AND WHAT TO DO ABOUT IT

SEO jobs are changing to include the distribution of content. They now focus more broadly on the discoverability of a website. Focus on methods that drive brand awareness growth. Pivot to tracking LLM visibility, structure content for LLM digestibility, change content approaches, and think about the future of conversion to prepare for the next few years.

THE SUMMER OF VIBE CODING IS OVER

Companies like Anysphere face soaring LLM inference costs, forcing price hikes and prompting exit strategies like reverse acqui-hires. The coding AI market's rapid growth, driven by reasoning models, is now pressured by expensive compute costs, leading vendors to adopt usage-based pricing.

AND Digital. Join live on September 16

OPENAI REORGANIZES RESEARCH TEAM BEHIND CHATGPT'S PERSONALITY

Following several public incidents regarding ChatGPT's personality, OpenAI is absorbing the Model Behavior team into the larger Post Training group, suggesting that personality is now considered a core capability to measure and hone during model development.

GPT-5 THINKING IN CHATGPT (AKA RESEARCH GOBLIN) IS SHOCKINGLY GOOD AT SEARCH

GPT-5 Thinking leads to far more comprehensive results, but it's a lot slower.

ANTHROPIC'S LEADING RESEARCHERS ACTED AS MODERATE ACCELERATIONISTS

Anthropic's researchers have embraced moderate accelerationism, advocating for measured and responsible AI advancement.

MEDICARE WILL START USING AI TO HELP MAKE COVERAGE DECISIONS NEXT YEAR

Starting January, Medicare will implement an AI pilot program in six states to assist in making coverage decisions through the WISeR Model.

KIMI K2-INSTRUCT-0905 RELEASE

Moonshot AI has released Kimi-K2-Instruct-0905, a mixture-of-experts (MoE) model featuring 32B activated parameters out of 1T total.

BITCOIN STAYS BELOW $112K AFTER TOUGH JOBS REPORT AND FED CUT BETS

The August US jobs report revealed only 22,000 job additions versus 75,000 expected, pushing Fed rate cut probability to 100% for the September 17 meeting, yet Bitcoin failed to capitalize on this typically bullish macro development and remains below $112,000. Technical analysis shows BTC has confirmed a bearish double-top breakdown below the $111,982 neckline, with the next support level at $101,700 around the 200-day moving average. Despite impending rate cuts, analysts warn that Treasury yield volatility could increase, potentially creating financial tightening conditions similar to late 2024 when the 10-year yield actually rose from 3.6% to 4.8% even as the Fed cut rates. The upcoming August CPI data next week is expected to show continued inflation stickiness at around 2.9% year-over-year, which could limit the downside in Treasury yields and maintain pressure on risk assets like Bitcoin.

SENATE DRAFT BILL PROPOSES SEC-CFTC JOINT COMMITTEE, DEFI PROTECTIONS

The latest draft of the Responsible Financial Innovation Act calls for a joint SEC-CFTC advisory committee to clarify crypto regulation and its treatment of tokenized RWAs. The bill explicitly protects DeFi developers and users from being treated as broker-dealers for running DEXs, validators, or wallets; exempts Decentralized Physical Infrastructure Networks (DePINs) from securities laws if sufficiently decentralized; and clarifies that staking rewards and airdrops are not securities offerings. Lawmakers aim to reconcile the bill with the House's Clarity Act and put it on President Trump's desk by year-end.

JUSTIN DRAKE'S “LEAN ETHEREUM” VISION

Lean Ethereum is a technical roadmap that succeeds the Beam Chain project and aims to achieve 10,000+ TPS on Ethereum's L1 while maintaining full decentralization and chain verification from mobile devices or browsers. The vision relies on two key technical breakthroughs: real-time zkVMs that can prove EVM execution in seconds for under a cent per transaction, and data availability sampling (DAS), which allows lightweight clients to verify massive blob throughput without downloading full datasets. Lean Ethereum proposes comprehensive upgrades across all three layers - execution (EVM 2.0 with SNARK-friendly instruction sets), data (blobs 2.0 with DAS and post-quantum cryptography), and consensus (beacon chain 2.0 with fast finality) – unified by hash-based cryptography as the foundational building block. If successful, the roadmap could position Ethereum as the dominant "internet of value" and settlement layer, capable of enterprise-scale throughput while preserving the security and decentralization properties that distinguish it from other Layer 1 networks.

HOW CRYPTO COULD TAKE OVER REAL WORLD PAYMENTS

FreePay is an open-source platform that enables in-person crypto payments through NFC tap technology. It is designed to work with any wallet, blockchain, or token without requiring proprietary hardware or apps. The system automatically selects the optimal payment token when customers tap their phones to merchant terminals, eliminating the interoperability issues that have plagued previous crypto payment attempts. FreePay provides merchants with the benefits of cash while maintaining global accessibility by offering instant settlement, zero chargebacks, and significantly lower fees compared to traditional credit card processing.

ON TEMPO, PERMISSIONLESSNESS, L1 VS L2

Stripe's Tempo blockchain will be a fully permissionless L1 with open token deployment, transactions, and eventual validation, despite targeting institutional payment use cases. The team chose to build an L1 rather than an Ethereum L2 because global payment partnerships require a decentralized validator set for trust and neutrality, eliminating dependencies on any single platform or sequencer. Technical priorities include fast finality, multiple validators, custom transaction lanes, and stablecoin-focused architecture with native issuance rather than Ethereum bridge functionality. The approach represents a bet that mainstream crypto adoption requires truly decentralized infrastructure rather than permissioned systems, aiming to onboard traditional payment flows onto permissionless crypto rails.

ETHEREUM RESEARCHERS PUSH FOR FASTER FINALITY

Ongoing Ethereum research to speed up transaction finality includes shortening slot times, exposing a fast confirmation rule that secures transactions in 1-3 blocks (10-30 seconds), and refining Gasper/beacon chain fork choice. Teams are also exploring new consensus mechanisms, such as three-slot-finality and fast BFT protocols, to cut down Ethereum's current 13 to 19 minute finality window. The goal is to bring faster confirmations to apps and infrastructure while maintaining Ethereum's security guarantees.

WHY BUSINESSES FIND CRYPTO USEFUL

Businesses are increasingly adopting crypto solutions. Stripe's new L1 blockchain Tempo will be a behind-the-scenes infrastructure similar to SWIFT or ACH rather than consumer-facing technology. Critical business drivers include instant onchain transfers that eliminate trapped liquidity during traditional banking delays, lower transaction fees compared to card payments, improved reliability for cross-border money movement, and reduced currency conversion requirements. Stablecoins provide access to USD-based financial functionality. Regulatory compliance, rather than arbitrage, is now the primary framework, given formal regulation under the GENIUS Act in the US and MiCA in Europe. The focus represents a shift toward solving practical enterprise payment challenges rather than speculative crypto use cases.

PAXOS, FRAX, AGORA COMPETE TO ISSUE HYPERLIQUID'S USDH STABLECOIN

Hyperliquid is preparing to launch USDH, a Hyperliquid-native stablecoin, with issuers competing for the contract through validator vote. Paxos proposed a MiCA- and GENIUS-compliant USDH with HYPE buybacks and zero-fee conversion from USDC, Frax offered a “community-first” model that forwards 100% of Treasury yields to users via frxUSD, and Agora pledged 100% net revenue share backed by State Street and VanEck reserves. A fourth team, Native Markets, also submitted a bid using Stripe-owned Bridge, while Ethena hinted it may join the race. Validator voting will determine the issuer on September 14.

ETH STABLECOIN SUPPLY ALL-TIME HIGH

The Stablecoin supply on Ethereum has reached a new all-time high of $172.2 billion.

POLYMARKET SETS RECORD FOR NEW MARKETS AHEAD OF U.S. RETURN

Polymarket created 13,800 new markets in August, its highest monthly total ever and 2,000 above July's record, as it prepares to re-enter the U.S. following CFTC approval via QCEX.