In partnership with
Welcome, AI enthusiasts
Anthropic has unveiled Anthropic’s Claude Opus 4.7, its most powerful publicly available model yet, with clear upgrades in coding and real-world task performance. Built-in limits and safeguards keep its more sensitive cyber abilities in check. Meanwhile, the even stronger Mythos Preview remains under wraps.. for now. Let’s dive in!
In today’s insights:
Anthropic Launches Claude Opus 4.7 While Mythos Stays Locked Away
Codex Becomes OpenAI's Always-On Coding Agent
Google's AI Mode Opens the Web Side-by-Side
Read time: 4 minutes
LATEST DEVELOPMENTS
Evolving AI: Anthropic released Claude Opus 4.7, a more capable public model, while keeping its stronger Mythos Preview restricted.
Key Points:
Opus 4.7 is Anthropic's most powerful generally available model, with gains in coding, instruction following, and real-world work.
Its cyber capabilities were deliberately limited, and automated safeguards block high-risk cybersecurity requests.
Mythos Preview remains restricted to select partners under Project Glasswing, and Anthropic calls it their best-aligned model to date.
Details:
Opus 4.7 retook the top spot for agentic coding among publicly available models, scoring 64.3% on SWE-bench Pro and SWE-bench Verified. It arrives two months after Opus 4.6, continuing a steady upgrade cadence, and reportedly pulls ahead of GPT-5.4 and Gemini 3.1 Pro on key benchmarks. Pricing stays at $5 per million input tokens and $25 per million output tokens, and the model is live across Claude products, the API, Bedrock, Vertex AI, and Microsoft Foundry.
Why It Matters:
There's a real shift happening here. A year ago every lab raced to ship their most capable model the moment it was ready. Now Anthropic is shipping the second-best one on purpose and openly saying the top one found thousands of zero-day vulnerabilities in testing, so only 50 banks and partners get to touch it. With California's SB 53 live and the EU's Code of Practice hitting enforcement in August, expect more "we built it but you can't have it" launches. The frontier is splitting into a public tier and a vault tier, and where your model sits is becoming a policy decision, not a product one.
The ones showing up in LLMs convert 3× better than Google
They optimized for LLMs, not just Google.
FAQs. Comparison pages. Transparent pricing. LinkedIn presence. These aren't vanity plays. They're what gets you cited in ChatGPT, Gemini, and Claude when your buyers are researching, your investors are looking, and your future hires are deciding where to work.
Download the free AEO Playbook for Startups from HubSpot and get the exact checklist. Five minutes to read.
Evolving AI: OpenAI's Codex can now control a Mac, run for weeks, and rival Claude Code head-on.
Key Points:
Codex gains "background computer use," letting it see the screen, move a cursor, and type across any macOS app.
The agent can schedule its own future tasks and keep working on projects across days or even weeks.
Over 90 new plugins connect Codex to JIRA, GitLab, Slack, Microsoft Suite, and more.
Details:
OpenAI has rolled out a major Codex update that turns it from a terminal helper into a full desktop agent. The new background computer use feature lets Codex operate any Mac app by seeing, clicking, and typing with its own cursor, with multiple agents running in parallel. A built-in browser allows developers to leave comments directly on web pages as instructions. Codex now also uses gpt-image-1.5 for visuals and can resume old conversation threads to keep context alive.
Why It Matters:
Coding assistants have quietly turned into coworkers you leave running overnight. Roughly 85% of developers already lean on AI day to day, Claude Code is pulling in around 2.5 billion dollars a year, and now Codex wants its own seat at the desk with a cursor that never sleeps. The shift is less about speed and more about trust. When an agent is clicking around your machine for days, reviewing its work stops being optional and starts being the actual job. So are we still writing software, or just signing off on it?
THE AI BRIEF
⚡️ Introducing AI Brief - Evolving AI's Quarterly Workshop
Turn AI into something you actually use in your work every day →
This is the first time we are doing a workshop, and we didn’t build it to be another webinar you half-watch from a second tab. AI Brief is a live, 6-hour quarterly workshop where we take everything that moved in AI this quarter and turn it into real workflows and use cases.
Here's what's inside:
Every major model, tool, and shift from Q1 2026, curated, explained, and made useful
Role-specific updates curated for PMs, founders, engineers, marketers, and designers
A hands-on build session where you walk away with at least 3 workflows ready to use the next day
The Q1 AI Brief Toolkit, including a prompt library and workflow templates that you keep
Access to a private community to ask questions, network, and grow
As part of our newsletter community, enjoy $50 off with code EARLYBIRD-NEWSLETTER at checkout.
Limited seats left.
Evolving AI: Google's AI Mode can now show webpages alongside its conversational search results in Chrome desktop.
Key Points:
Clicking a link in AI Mode now opens the webpage next to the chat, so users can compare details and ask follow-up questions without losing context.
A new "plus" menu lets users pull open Chrome tabs, images, and files directly into an AI Mode search.
The features are live in the U.S., with expansion to other regions planned.
Details:
Google rolled out the update on April 16, letting Chrome desktop users keep AI Mode open while browsing a linked page. The AI pulls context from the visible webpage and the wider web to answer follow-ups. If you are shopping for a coffee maker, you can ask how easy a specific model is to clean and get an answer grounded in the retailer's page. On desktop and mobile, the new tab-mixing feature lets users combine notes, slides, and open tabs into one search, useful for research or studying.
Why It Matters:
Publishers are already bleeding. Google search referrals to publishers dropped about a third in 2025, and small sites lost closer to 60 percent. Side-by-side browsing sounds user-friendly, but it keeps people parked inside Google while the source page becomes just a reference tab. With roughly 58 percent of searches now ending in zero clicks, this update quietly nudges that number higher. The open web still runs on people actually visiting websites, and every design choice like this one chips away at that foundation.
This video from Google explains quantum mechanics in a simple way, showing how particles can exist in multiple states and stay connected across distance. These same principles are the foundation of quantum computing, which could massively accelerate AI and unlock problems today’s systems can’t solve. Still early, but this is where things get interesting.
👀 Click on the image you think is real
QUICK HITS
🖥️ Perplexity realeased Personal Computer.
🤖 OpenAI Launches GPT-Rosalind: Its First Life Sciences AI Model Built to Accelerate Drug Discovery and Genomics Research.
📈 Struggling shoe retailer Allbirds makes bizarre pivot to AI, adds $127 million in value.
🌐 NVIDIA Launches Ising, the World’s First Open AI Models to Accelerate the Path to Useful Quantum Computers.
🖥️ The Gemini app is now on Mac.
📈 Trending AI Tools
🗣️ Wispr Flow - Voice-to-text AI that turns speech into clear, polished writing in every app*
📄 Kickresume - AI resume builder
✍️ Evernote - Turn audio, video and images to text
🎨 Beautiful AI - A faster way to create professional presentations
🧠 Brainly - AI learning companion
*partner link