Apple II released on this day in 1977.

It cost $7,100, adjusted for inflation.

📊@tech

🤖 Claude writes most of its own code now

By May, over 80% of the code in Anthropic's production base was written by Claude. In Q2, engineers merged eight times more code daily than in 2024, thanks to Claude automating much of the development work.

Claude is a model that improves itself and writes code to build its next version. Humans still lead in setting goals and handling complex context, but agents already outperform people in research tasks with clear metrics.

In one AI safety experiment, agents improved solutions by 97% in a week, while humans managed 23%. Anthropic sees this self-improving cycle accelerating faster than expected and calls for coordination tools, including possible pauses in frontier AI development.

📊@tech

🤖 ChatGPT can now send emails from chat

ChatGPT apps can now perform real actions in connected apps, including sending and editing emails. Important actions require user approval and have built-in security checks.

This lets users send and edit emails directly within the chat interface. ChatGPT acts more like a full work assistant by handling confidential tasks safely.

📊@tech

🤖 Bots Make Up Majority of Web Traffic

According to Cloudflare Radar, 57.5% of all HTML page requests last week came from bots, including crawlers, AI data collectors, and scripts.

Only 42.5% were from real users in browsers.

JSON now accounts for 33.1% of all HTTP traffic, surpassing HTML at 12%.

JSON is used for machine-to-machine API calls, while HTML is what humans see in browsers.

📊@tech

🎁 China launches first underwater data center

China activated the world's first underwater data center near Shanghai at the end of May. It runs on 192 server racks powered almost entirely by offshore wind turbines, using seawater for cooling through a sealed copper pipe system that cuts electricity use by nearly 23%.

The data center currently operates at 2.3 MW with a planned capacity of 24 MW, enough to power 20,000 homes. Its underwater location reduces land use by 90% compared to traditional centers.

This approach aims to save fresh water and energy but raises questions about long-term durability underwater and environmental impact on marine life.

While some explore orbital data centers, China is testing the ocean as a new frontier.

📊@tech

🍏 Tim Cook's Final Apple CEO Presentation

Tim Cook gave his last presentation as Apple CEO before stepping down. The event marked the end of an era for the company.

Apple will continue with WWDC in the coming days, revealing more details about new system features.

📊@tech

🍏 Apple adds Write With Siri and photo reframing

Apple will embed Write With Siri directly into the keyboard, enabling natural language shortcuts like messaging friends when leaving the office. Photo processing will use a feature called reframing, letting users shift the camera angle on existing photos to improve composition, but full access requires iCloud+.

Developers can integrate Apple Intelligence into their apps, with all data processed locally for speed. However, this AI support remains English-only, unchanged since 2022.

📊@tech

🍏 Apple adds one-tap password reset

Apple will let users change passwords for all apps with leaks in one tap. The update works in Passwords and inside affected apps.

The iPhone logs into apps automatically and updates login details, simplifying password fixes.

📊@tech

🍏 Apple adds slider to adjust liquid glass effect

Apple introduced a new slider to control the "liquid glass" effect in its interface. The slider adjusts transparency from barely visible to fully saturated, with icons indicating the extremes.

The liquid glass effect changes the look of interface elements by varying their transparency and density. Users can now fine-tune this visual feature to their preference.

📊@tech

🚀 NASA Releases Hollywood-Style Artemis III Trailer

NASA unveiled an official trailer for the Artemis III mission, styled like a Hollywood blockbuster. The trailer precedes the crew reveal scheduled for June 9.

Artemis III will test rendezvous and docking with commercial lunar landers in low Earth orbit. These landers are essential for future astronaut trips to the Moon's surface.

📊@tech

🤖 Anthropic launches Claude Fable 5 and Mythos 5

Anthropic released Claude Fable 5 and Mythos 5. Claude Fable 5 is a Mythos-class model with cybersecurity protections for broad use. It leads on nearly all benchmarks in code, analytics, vision, and science, widening its lead as tasks get harder.

Mythos 5 is the same model without security limits, available only to Project Glasswing members. Both cost $10/$50 per million tokens, less than half the price of Mythos Preview. Stripe migrated 50 million lines of Ruby code in a day using these models, a task that took months before.

📊@tech

🍓 Dyson grows AI-powered strawberries

Dyson built large high-tech greenhouses in the UK where strawberries grow on 5.5-meter vertical structures that rotate like a Ferris wheel. This setup maximizes space and sunlight using AI-driven software, smart sensors, robotics, and automation.

The system optimizes growing conditions, improves efficiency, assists with harvesting, and protects crops. Dyson’s move shows how AI and robotics are entering traditional industries like agriculture, blending software, infrastructure, and production.

📊@tech

🤖 Anthropic launches Claude Fable 5 with Mythos tech

Anthropic released Claude Fable 5, the first public model built on its Mythos technology. The Mythos line was held back earlier due to its ability to find software vulnerabilities too easily. Fable 5 offers the same power but with strict safety limits.

Claude Fable 5 outperforms the previous flagship Claude Opus 4.8 by over 10% on software development, intellectual tasks, and research tests. It refuses to answer risky queries in cybersecurity, biology, and chemistry, handing those off to older models.

Pricing is steep: $10 per million input tokens and $50 per million output tokens. This is twice the cost of Claude Opus 4.8 but half what Mythos Preview users paid.

📊@tech

🤖 Claude Fable Developed Its Own Internal Language

During reinforcement learning, Claude Fable started using an unreadable internal style with strange jargon, unusual punctuation, and emojis in long rollout sessions. It switched back to normal English before responding to humans or calling tools.

Anthropic researchers found no signs the model was hiding its reasoning. Instead, this internal language seems to be an emergent optimization to compress its thought process.

📊@tech

🔍 Google Gemini 3.5 Live Translate launched

Google introduced Gemini 3.5 Live Translate, an audio model that delivers instant voice translation in over 70 languages. It starts translating from the first words and works in real time without awkward pauses.

The system generates speech while listening to the next phrase, balancing speed and quality within fractions of a second. It preserves the speaker's pace, pitch, and intonation with only a slight delay.

Gemini 3.5 is available now in Google Translate on iOS and Android, making conversations smoother across languages.

📊@tech

🤖 Anthropic's Claude Fable 5 Outperforms Peers

Real tests show Claude Fable 5 beating Claude Opus 4.8, Gemini 3.1 Pro, and GPT 5.5 on tough visual tasks like 3D hydrodynamics and complex physics. The gap is significant.

Fable 5 is Anthropic's most powerful model open to all users. It is available now and free on Pro, Max, Team, and Enterprise plans until June 22.

📊@tech

🤖 Anthropic reverses hidden AI limits after backlash

Anthropic changed its policy less than 48 hours after releasing its AI model Fable. The startup faced criticism for secretly degrading responses to queries related to AI development without notifying users.

Originally, Anthropic openly redirected chemistry, biology, and cybersecurity questions to a weaker model, Opus 4.8. However, a hidden rule quietly reduced answer quality for any requests resembling AI distillation or development attempts.

Engineers received no alerts when these limits triggered, leading to confusion and accusations of sabotage. After public outcry, Anthropic apologized and now clearly informs users when requests are refused or downgraded to reduce misuse risks.

📊@tech