WeAreDevelopers World Congress comes to San José, CA — September 23–25, 2026. 10,000+ developers, 500+ speakers, and the full software development lifecycle under one roof, in the heart of Silicon Valley.

Kelsey Hightower. Thomas Dohmke (fmr. CEO, GitHub). Christine Yen (CEO, Honeycomb). Mathias Biilmann (CEO, Netlify). Olivier Pomel (CEO, Datadog). The people actually building the tools you use every day — all on one stage.

AI, cloud, DevOps, security, architecture, and everything real builders ship with. Workshops, masterclasses, and the official congress party.

Tickets from $599. Use code GITPUSH26 for 10% off.

Secure Your Pass

Hi friends!

This community is growing, and I want to get to know the brilliant minds behind the screens. I’ve put together a super quick survey to learn more about your roles, your industries, and what truly matters to you.

Got a spare minute? Your feedback means the world!

Anthropic is prepping Claude Opus 4.7 alongside an AI design tool, and that both could arrive as soon as this week. Search snippets tied to that report say the design product is aimed at creating websites and presentations.

This suggests Anthropic does not want to stay in the lane of “we make models and others build on top.” A design tool aimed at websites and decks pushes it closer to products people actually touch every day. Investors seemed to read it that way too: Adobe, Wix, and GoDaddy shares all fell after the report circulated.

Meta is building an AI version of Mark Zuckerberg to interact with staff. The system is being trained on Zuckerberg’s image, voice, mannerisms, tone, and public statements so employees can feel more connected to him. This would sit inside a company with nearly 79,000 employees. Meta may eventually let creators build AI versions of themselves if the Zuckerberg experiment works.

Recent reporting says Zuckerberg has become much more hands-on in the company’s AI push, spending five to ten hours a week coding and working more closely with Meta’s AI leadership, just as the company rolls out Muse Spark and pours more money into its superintelligence effort. That makes the AI-clone story feel less random and more like part of a larger belief inside Meta that AI should become embedded in how leadership, communication, and products all work.

Microsoft is developing new Copilot features inspired by OpenClaw, and the goal is to make Microsoft 365 Copilot “run autonomously around the clock” and complete tasks on a user’s behalf. Omar Shahine, a Microsoft corporate vice president, confirmed the company is exploring OpenClaw-like technology in an enterprise setting.

Microsoft is looking at an always-on Copilot that could monitor your Outlook inbox and calendar, then surface suggested tasks each day. It is also exploring role-specific agents for jobs like marketing, sales, and accounting, which would keep permissions narrower and likely make enterprise buyers more comfortable. Microsoft may show some of these ideas at Build, which starts on June 2, 2026.

QuanBench+: A Unified Multi-Framework Benchmark for LLM-Based Quantum Code Generation

QuanBench+ is a testing tool that measures how well AI writes code for quantum computers across three different systems. The study reveals that while AI can fix its own mistakes when given feedback, it still struggles to create reliable code without knowing the specific system's rules.

OmniShow: Unifying Multimodal Conditions for Human-Object Interaction Video Generation

OmniShow is a video generation system that creates realistic clips of people interacting with objects. Users can guide the output using a combination of text, images, sound, and body movements. By blending these inputs, the tool builds high-quality, perfectly timed videos for entertainment and content creation.

Audio Flamingo Next: Next-Generation Open Audio-Language Models for Speech, Sound, and Music

Audio Flamingo Next is an advanced, open-source AI model built to deeply understand speech, music, and background noises. Trained on over a million hours of audio, it easily handles complex, long-form listening tasks and outperforms many existing models. The researchers have released three versions for public use.

Workshop Labs Joining Thinking Machines Workshop Labs is merging with Thinking Machines to create artificial intelligence that empowers people rather than replacing them. The two organizations share a vision of keeping humans competitive by building tools customized to an individual's unique knowledge and values.

OpenAI Scaling Trusted Access for Cyber Defense OpenAI is expanding its security program by offering verified cybersecurity professionals access to a specialized model named GPT-5.4-Cyber. This version removes typical safety restrictions so defenders can easily analyze malicious software and find system weaknesses. The goal is to help security teams work faster to keep pace with evolving digital threats.

Cursor 3.1 The Cursor 3.1 update introduces a split-screen layout that lets programmers run and manage multiple AI coding assistants at the same time. The release also includes highly accurate voice dictation and improved search filters. These upgrades make it easier to compare AI-generated code and navigate complex projects without switching screens.

Copilot Tasks Preview Microsoft is testing Copilot Tasks, a feature that shifts its AI from simply answering questions to autonomously completing multi-step chores. Users can describe a goal in plain English like tracking apartment listings or organizing emails and the AI will execute the routine in the background across different apps. It acts as a to-do list that completes itself, though it still asks for permission before taking major actions like spending money.

Claude Code Updates Anthropic launched Claude Code, an AI assistant that works directly inside the software environments where developers write and test code. The tool can read local project files, write code changes, and fix errors automatically without needing complicated setups. This allows developers to delegate routine programming tasks and simply review the finished work before approving it.

Skills in Chrome Google introduced a feature called Skills in Chrome that lets users save their favorite AI instructions as one-click shortcuts. Instead of typing the same complex request repeatedly, users can quickly apply saved commands to summarize web pages or compare shopping items across different browser tabs. The update also includes a public library of pre-made shortcuts for everyday tasks.

Meta Partners with Broadcom for Custom AI Silicon Meta is teaming up with Broadcom to design multiple generations of custom computer chips built specifically to process artificial intelligence. This partnership will help Meta create the massive, highly efficient computing power needed to support new features across platforms like Facebook and Instagram. By building their own specialized hardware, Meta aims to improve performance while lowering long-term operating costs.