The AI web browser that remembers your logins

Welcome back, Superhuman. When a launch video literally tosses its competition in the trash, the company behind it had better be confident in its product. That's the risk one startup just took with Aside, a new AI web browser that can access your logged-in accounts to get tasks done — without any hand holding.

Today: OpenAI upgrades GPT-5.5 Instant, how to chase overdue invoices with AI, and get the latest prompts and trending social posts.

1. Aside launches an AI browser that works across logged-in websites to get tasks done: The new agentic browser turns your browsing history into local on-device memory, then uses autofill to sign into your accounts without any help to complete tasks. The YC-backed browser claims the #1 spot on three browser-agent benchmarks, and the launch video shows the AI identifying and canceling unused subscriptions. Try Aside or watch the viral launch video (1M views).

2. OpenAI rolls out a new version of GPT-5.5 Instant, along with its first custom chip: The updated GPT-5.5 Instant — rolling out now — will better understand the intent behind your prompts or comments, making it more conversational and fun to talk to. The AI leader also unveiled Jalapeño, its first custom chip co-developed with Broadcom and designed to run inference for OpenAI’s products and help reduce its dependence on Nvidia. Learn more about Jalapeño.

3. Google poised to lose two more top researchers: Bloomberg reports that Jonas Adler and Alexander Pritzel — both key contributors to Google's flagship Gemini model — are preparing to join Anthropic. The two departures would continue a talent exodus from Google, with Nobel laureate John Jumper also leaving for Anthropic and top researcher Noam Shazeer joining OpenAI.

Join leaders from Prudential Insurance, Siemens, GAF, and HF Sinclair to learn how to build resilient data foundations for enterprise AI.

Get practical strategies from data and AI experts on trusted data products, intelligent agents, and strong governance frameworks in this virtual panel.

Register now.

Europe 2031. Eight researchers published Europe 2031, a thought piece calling for “the most ambitious political agenda” in the history of post-war Europe. Their argument: the EU has fallen behind in the AI race and the window to catch up is narrowing. Without action, they warn, Europe will lose the power to shape its own future — not over the next decade, but within five years.

Sliding into irrelevance. The essay argues European leaders have repeatedly misread the speed and scale of the AI race, treating it with complacency and skepticism instead of action. As a result, the EU controls few homegrown frontier models and just 5% of global compute compared to America’s 80%. Europe has almost no leverage, and the researchers venture a guess as to how the future might unfold from here.

Looking ahead. By 2031, Europe (theoretically) could find itself in a position where the US owns the cognitive frontier, China leads in robotics, and everyone else is at a disadvantage. Europe would be forced to pick a side or — in the authors’ words — “wither away in isolation.”

Righting the ship. The researchers’ suggestions? Pour massive investments into compute, rally a coalition of other lagging nations to boost collective bargaining power, and position Europe to win the coming physical-AI revolution. The essay echoes Citrini Research’s The 2028 Global Intelligence Crisis, which explores a similar hypothetical future for AI’s impact on America’s economy.

Sample prompt: “@Viktor pull every invoice more than 14 days overdue from QuickBooks. Draft a friendly payment reminder for each client with the invoice number and amount. Post them here for approval before sending. Once approved, follow up every 3 days until each invoice is paid.”

You can set up Viktor in Slack and Teams here.

Your next enterprise deal will ask for SSO before it closes. Most teams scramble for weeks. With WorkOS, you ship SSO, SCIM, and audit logs in a sprint - the same stack Anthropic, OpenAI, Cursor, and Perplexity run on.

Start building free

🧑‍💻 Claude Tag: Andrej Karpathy says Claude Tag is a “new paradigm” for using Claude, but others aren’t so sure, with some claiming the update isn’t anything special. Read his full comment and discourse (4.5M views).

📝 Consent Registry: Want to prohibit AI tools from using your likeness? Consider adding your name to the Human Consent Registry, a new record where you can publicly post your preference.

⏳ Back in Time: AI has created a new genre of content: historical POV videos. Watch one modern-day vlogger explore a day in the life of 1536 London (5M views).

💿 2000s Style: Someone imagined what the cover art might look like for Claude, ChatGPT, Gemini, and Grok if they were CDs from the 2000s.

👀 Dead Internet: A Redditor noticed that a World of Warcraft server that has no real humans — just 1,800 DeepSeek bots. The post say it’s the dead internet theory in action, but many comments disagree (5.8K upvotes).

^{* indicates a promoted tool, if any}

Anthropic turned Claude into a teammate that lives inside Slack. You can now tag @Claude in any channel and the AI will jump in to help, pulling context from the existing conversation. Try it out or see how it works.

Grow customers & revenue: Join companies like Amazon, HubSpot, and Salesforce. Showcase your product to our 1M+ readers and 2M+ followers on socials. Get in touch.

Until next time — Zain, Theodore, & the Superhuman AI team.