Search the archive

FROM EXPERIMENTAL TOOL TO AI INFRASTRUCTURE: ADAM ZIELIŃSKI'S VISION FOR WORDPRESS PLAYGROUND IN 2026

Zieliński's goal for 2026 is to make Playground useful and easily accessible for a large crowd out there that's not super technical.

A UNIX MANIFESTO FOR THE AGE OF AI

The Unix philosophy has survived because its authors built restraint into the architecture.

I BUILT A PROGRAMMING LANGUAGE USING CLAUDE CODE

Cutlet is a dynamic programming language built entirely by Claude Code that combines the expressiveness of Python, Ruby, Lua, and JavaScript with first-class support for running subprocesses, building pipelines, and scripting.

Learn more about evals →

ANTHROPIC SUES PENTAGON OVER ‘SUPPLY CHAIN RISK' LABEL

Anthropic has filed two lawsuits against the Department of War to change the Pentagon's decision to label it a 'supply chain risk'. One of the lawsuits is in the US District court in the Northern District of California, and the other is in the US Court of Appeals for the District of Columbia Circuit. The supply chain risk designation is typically applied to firms that are deemed a major national security risk and has never been used on an American company. Anthropic has offered to continue negotiating with the Pentagon and also offered to help move the Pentagon off its technology and onto another AI system. OpenAI and xAI have signed agreements to provide technology on the Department of War's classified systems in recent weeks.

APPLE POSTPONES SMART HOME DISPLAY LAUNCH AS IT WAITS FOR NEW AI AND SIRI

Apple's smart home display has been delayed until later this year. The project was first scheduled to launch in the spring of 2025, but it was postponed to let the company finish work on a new Siri digital assistant. It was then scheduled to be released this month, but the new Siri is still not yet ready. The display is designed to be a central AI hub for the home that can display personalized data, such as calendar appointments, reminders, notes, and more.

CHINA LEADS THE HUMANOID ROBOT RACE — BUT THE US STILL HAS A SHOT

China's humanoid robots now dominate the market with over 90% of global sales and thousands of units shipped last year. Tesla's Optimus robots still won't be ready for launch until at least next year. While Chinese vendors are more advanced when it comes to production scales, US companies are very strong at the technical side of things, especially in the hardware and software departments. By the time humanoid robot startups build up their production bases, they will be ready for large-scale deployment.

HOW TO DESIGN ANTIBODIES

BoltzGen is the leading open-source approach for computational antibody design. It uses the permissive MIT license, so it can be used commercially by anyone. This guide walks through the full process of designing an antibody from home using BoltzGen. The process involves choosing a target, preparing a target structure, running a design campaign, filtering candidates, and experimentally validating the results.

Schedule a demo to learn more →

CODE REVIEW, A NEW FEATURE FOR CLAUDE CODE

Code Review is a new feature for Claude Code that dispatches a team of agents on every PR to catch bugs. Built for depth, not speed, the system is now in research preview for Team and Enterprise. Anthropic runs Code Review on nearly every PR. The tool doesn't approve PRs, but it closes the gap so reviewers can cover what's shipping. Reviews are billed on token usage, and admins have several ways to control spend and usage.

PERHAPS NOT BORING TECHNOLOGY AFTER ALL

The recurring concern, that large language models will push technology choices towards the tools best represented in their training data, making it harder for new tools to break through the noise, doesn't really hold up anymore. New models have large enough context lengths that they can consume a lot of documentation before they start working on a problem. Most agents work just fine in existing codebases that use libraries or tools too private or new to feature in the training data. Developers are still free to choose whatever tools they want to use and are not restricted to using the ones LLMs are most familiar with.

AFTER FALLING FAR BEHIND THE REST OF INDUSTRY, BLUE ORIGIN CREATES NEW STOCK OPTION PLAN

When Jeff Bezos launched Blue Origin, he knew that the company would not meet investors' expectations for return on investment over a typical investing horizon. Decades later, the company is still not operationally profitable, though recently, it has made impressive strides and seen financial returns from the sale of engines and commercial launches. To continue its growth and attract top talent, Blue Origin will begin granting stock options to employees this spring. The new program is structured to provide opportunities for liquidity events that will enable employees to convert vested stock options into realized value. More details about the program will be released during a company-wide meeting on April 17.

AMAZON TELLS FCC TO BIN SPACEX'S MILLION-SATELLITE DATACENTER DREAM

Amazon has criticized SpaceX's application for permission to launch a fleet of orbital datacenter satellites as incomplete, speculative, and unrealistic. It wants regulators to reject the application, which it says is a speculative placeholder rather than a complete application under the Commission's rules. Amazon also raised concerns about satellite interference and environmental objections. Analysts say that SpaceX's plan of putting datacenters in space is 'peak insanity' as running spaceborne facilities would be uneconomical and could never satisfy terrestrial demand for compute power.

FORMER META AI CHIEF'S START-UP IS VALUED AT $3.5 BILLION

Yann LeCun's Advanced Machine Intelligence Labs is only a month old and employs just 12 people.

GHOSTTY 1.3.0

Ghostty 1.3.0 is a significant release that includes hundreds of improvements, bug fixes, and performance optimizations across all platforms.

BLUESKY CEO JAY GRABER STEPS DOWN

Graber, who will be replaced by Toni Schneider as interim CEO, will transition to a new role as chief innovation officer.

VIDEO CONFERENCING WITH POSTGRES

SpacetimeDB recently open-sourced a way for people to make video calls over a database.

10X IS THE NEW FLOOR

AI amplifies people with agency and curiosity.

THE HUMAN.JSON PROTOCOL

human.json is a lightweight protocol for humans to assert authorship of their site content and vouch for the humanity of others.

Schedule a demo to learn more

CURSOR GOES TO WAR FOR AI CODING DOMINANCE

Until recently, Cursor seemed nearly unstoppable. Then, Anthropic's Claude Code crossed $1 billion in annualized revenue within six months and hit $2.5 billion last month, surpassing Cursor, and OpenAI's Codex coding agent was downloaded over a million times in its first week. Perceived momentum can appear or evaporate overnight in the fast-moving world of AI. In response, Cursor is building a research powerhouse to outmaneuver Anthropic and OpenAI, and prioritizing contracts with large enterprises, which can be more stable than consumer subscriptions.

APPLE ‘ULTRA' PRODUCTS EXPANSION IS UP NEXT AFTER MACBOOK NEO LAUNCH

Apple plans to release several higher-end products, including a foldable iPhone, new AirPods, and a MacBook Pro with a touch-enabled OLED display. Other product lines will likely expand to a superpremium tier, including the iPad and iMac. The approach is typical of CEO Tim Cook's playbook of taking proven products and stretching them across multiple price points to capture as much market share as possible. It wouldn't be surprising if the MacBook Neo eventually expanded into multiple variants, even a larger-screen model.

STEM CELL TREATMENTS FOR PARKINSON'S AND HEART FAILURE APPROVED IN WORLD FIRST

Japanese pharmaceutical company Sumitomo Pharma has received a conditional and time-limited approval for the manufacture and sale of Amchepry, a Parkinson's disease treatment. The treatment involves transplanting stem cells developed into the precursors of dopamine-producing brain cells into a patient's brain. Medical startup Cuorips has also received approval from Japan's health ministry to manufacture and sell ReHeart, heart muscle sheets that can help form new blood vessels and restore heart function. The treatments could be on the market and rolled out to patients in Japan as early as this summer.

ASTEROID DEFENSE MISSION SHIFTED THE ORBIT OF MORE THAN ITS TARGET

The Double Asteroid Redirection Test (DART) operation in September 2022 changed the trajectory of the entire Didymos binary system and altered its orbit around the Sun. The vending-machine-sized probe crashed into the smaller asteroid at over 22,000 kilometers per hour, decreasing the along-track velocity of the entire system by roughly 11.7 micrometers per second. Some of the rock and dust created by the impact gave the asteroids a push roughly equal to the initial impact of the spacecraft itself, further changing the asteroids' orbits. The final verification of the DART mission's consequences will come in late 2026 when the ESA's Hera spacecraft arrives at the Didymos system.

Upgrade to Core 3 today

I DON'T KNOW IF MY JOB WILL STILL EXIST IN TEN YEARS

It is unclear if the software engineering industry will survive another decade. If it does, it will certainly change more than it did in the last two decades. Software allows for the automation of jobs, which is why it is such a lucrative profession. There aren't any genuinely new capabilities that AI agents need in order to automate software engineering, they just have to get better and more reliable at doing the things they can already do. It's hard to believe that AI will cause demand for software engineers to increase rather than decrease.

THE LINUX KERNEL WILL SOON BE MIT-LICENSED AND COPYLEFT WILL BE DEAD WITHIN 5 YEARS

A disagreement in a project licensed under the GPL resulted in a developer reimplementing the project with AI from start to finish. The AI generated a complete clean room implementation with just 1.3% in common code. This scenario signals the end of copyleft. Companies that want to use GPL software will likely rather spend a week reimplementing projects with AI assistance than dealing with the headaches of licensing.

STABLECOIN FIRMS BET BIG ON AI AGENT PAYMENTS THAT BARELY EXIST

Circle and Stripe are racing to build a payments system for a world where autonomous AI agents transact millions of times a day using stablecoins. Circle's Arc is a blockchain for stablecoin payments that allows autonomous agents to hold a balance and spend across networks with transaction costs of only fractions of a penny. Stripe's Tempo is a blockchain with a $5 billion valuation designed specifically for stablecoin payments. While fast and cheap, stablecoin transactions lack the fraud protection, dispute resolution, and credit extension that cards bundle into every payment. A potential path forward may be virtual cards that settle on the back end via stablecoins, which allows card networks and stablecoin rails to coexist rather than compete.

FOR OPENAI AND ANTHROPIC, THE COMPETITION IS DEEPLY PERSONAL

OpenAI created the fastest-growing consumer app in tech history. However, Anthropic managed to add thousands of big businesses as customers in just a few months and expects to see $19 billion in revenue this year. The recent controversy with the Pentagon has helped Anthropic, and its app has soared to the number one spot on Apple's App Store. The situation is just the latest round in a long-running and deeply personal feud between the two companies, whose executives have differing views on how AI should be created.

A MODEL OF HOW SIMPLICITY GETS REWARDED

You can reward simplicity, you just have to do it indirectly.

BUNDLE OF HUMAN NEURONS HOOKED TO SILICON LEARNS TO STUMBLE THROUGH DOOM

Roughly 200,000 living human neurons grown on a microelectrode array gradually learned to play Doom in response to feedback, a form of reinforcement learning.

TESLA OPENS ITS FIRST MEGACHARGER STATION TO SEMI CUSTOMERS IN CALIFORNIA

The Megacharger station at 4265 E Guasti Road in Ontario, California, is now live and available to Semi fleet customers.

DO AI-ENABLED COMPANIES NEED FEWER PEOPLE?

Claims that companies can get by with way fewer people in the age of AI appear to be true.

GNU AND THE AI REIMPLEMENTATIONS

AI reimplementations are a fair process, but developers should at least put some effort into trying to evolve the work when replicating it.

THE NEW APPLE FINALLY BEGINS TO EMERGE

Apple has added two designers, Steve Lemay and Molly Anderson, to its executive leadership page, which, when combined with the company's recent MacBook Neo launch, signals a turning point for the company.

2 certified. Your data never trains models. Watch Viktor work ->

Backed by Daniel Gross (Y Combinator), Nat Friedman (ex-GitHub CEO),

and Mati Staniszewski (ElevenLabs). See why ->

Give Viktor your worst task (for free) ->

INTRODUCING GPT‑5.4

OpenAI has released GPT-5.4. It is available as two new API models, gpt-5.4 and gpt-5.4-pro, and also in ChatGPT and Codex CLI. The model has a knowledge cut of date of August 31, 2025, and a one-million-token context window. OpenAI put a particular focus on improving the model's ability to create and edit spreadsheets, presentations, and documents. It beats GPT-5.3-Codex on all relevant benchmarks.

MS EXEC: MICROSOFT'S NEXT CONSOLE WILL PLAY “XBOX AND PC GAMES”

Microsoft's Executive Vice President for Gaming, Asha Sharma, says that the next-generation of Xbox console will play both Xbox and PC games. It is possible that the console's access to PC games could be limited to Microsoft's existing streaming solution via PC Game Pass, or to games designed for Microsoft's own Xbox-branded PC SDC and the PC Xbox app. However, this could also mean that Microsoft is getting ready to open up its next console to a complete Windows installation. The move could be a response to Valve's upcoming Steam Machine, which is threatening to bring Windows-free PC gaming to living rooms everywhere in the near future.

BRAIN TECH STARTUP SCIENCE CORP. RAISES $230 MILLION TO TREAT BLINDNESS

Neurotechnology company Science Corp. has raised $230 million, bringing its valuation to $1.5 billion. The funds will be used to commercialize its implant for blindness and develop even more advanced brain devices. Science Corp. is now the second-most valuable brain implant company after Neuralink. Its retina implant, PRIMA, is a chip that sits at the back of the eye to help blind people see with the assistance of a special pair of glasses that projects images into the eye. It has been demonstrated to improve vision for patients with advanced age-related macular degeneration. The device is currently under regulatory review in Europe and the US.

CONGRESS EXTENDS ISS AND TELLS NASA TO GET MOVING ON PRIVATE SPACE STATIONS

Senator Ted Cruz has introduced a new bill to step up the pressure on NASA to create competition among private companies to develop replacements for the International Space Station. The bill mandates that NASA publicly release requirements for commercial space stations in low-Earth orbit within 60 days, release the final request for proposals to solicit industry responses within 90 days, and enter into contracts with two or more commercial providers for such stations within 180 days. The legislation extends the International Space Station's lifespan from 2030 to 2032, but an extension must still be approved by international partners.

A GITHUB ISSUE TITLE COMPROMISED 4,000 DEVELOPER MACHINES

A version of Cline published to npm on February 17 was byte-identical to the previous version. However, it had a one-line change that installed OpenClaw onto users' machines. This resulted in approximately 4,000 OpenClaw downloads before the package was pulled. The attacker got the npm token by injecting a prompt into a GitHub issue title, which an AI triage bot read, interpreted as an instruction, and executed.

NEW "AUTO MODE" PERMISSIONS COMING...

Claude Code plans to launch Auto Mode in research preview no earlier than March 11. Auto Mode lets Claude handle permission decisions during coding sessions so developers can run longer tasks without being interrupted for manual approvals. It is designed to be a safer alternative to bypassing permissions entirely. Users should only use Auto Mode in isolated environments, as it isn't perfect and won't catch every action that could be considered risky.