Date: {{current_date_full_with_day}}

Hey {{first_name | AI Visionaries}},

In this 106th edition we talk about Cursor's report on how AI coding is changing developer work and the insights will surprise you.

We also talk the rise of one-founder AI-native companies aided by a very well funded startup called Polsia.

Thank you for being an engaged reader and helping us cross 5,500 subscribers over 106 editions. There are another 2,500 in our LinkedIn edition, so that is heartening to note.

As usual, please send in your feedback and suggestions - it helps me to improve the newsletter for you.

-Renjit

PS: If you want to unleash the power of AI agents for personal and business productivity, you can setup time speak to me here»

Set up time to discuss:

Anthropic released Claude Opus 4.8, a same-price upgrade that focuses less on flashy new tricks and more on making long, agentic work steadier.

The company says the model improves on Opus 4.7 across coding, reasoning, tool use, and practical knowledge-work benchmarks, with early users calling out better judgment, stronger browser and computer-use performance, and more reliable follow-through on complex tasks.

The launch also came with product changes that matter for teams building with AI. Claude users can now control how much effort the model spends on a task.

Claude Code added dynamic workflows for large-scale problems, and fast mode for Opus 4.8 runs at 2.5x speed while costing three times less than fast mode on earlier Opus models.

Anthropic also highlighted honesty improvements, including a reduced tendency to over-claim progress when the model has not actually completed the work. Laziness no more!

SO WHAT?

For founders, this is another sign that AI advantage is shifting from raw model quality to reliable agents that can finish messy, multi-step work. Let me know if you are building in this space.

Read the source:

Last week, Viktor wrote a campaign brief, built a landing page, opened a pull request, generated a board-ready PDF from live Stripe data, and sent a follow-up email to a churned customer. All from Slack. Same colleague that also pulls your reports and monitors your dashboards. 5,700+ teams. 3,000+ integrations.

Start free. $100 in credits →

Cursor published its first Developer Habits Report, using product data to show how quickly AI-assisted software development is moving from autocomplete into deeper agentic workflows.

The headline finding is simple: developers are adding more code, pull requests are getting larger, and agent sessions are becoming more complex as tools read files, edit code, run commands, and search more context on behalf of users.

The report says lines added per pull request are up roughly 2.5x year over year, average tool calls per agent session rose about 30% in two months, and accepted AI-generated lines are surviving longer after review.

It also found a wide cost gap between model families, with cost per agent request varying by nearly 9x, and a power-user gap where the most advanced developers get a much larger lift from AI than casual users. Better developers using better tools yield better results!

SO WHAT?

The practical lesson is that simply buying AI tools is not enough; companies need workflows, review habits, and power users who know how to turn agents into real shipping speed. So we go back to studying Computer Science, again?

Read the source:

Morning Brew unpacks the biggest stories in business, finance, and tech every morning.

Explained clearly. Read quickly. And a little fun too.

Try it yourself and join 4 million+ professionals reading daily.

Check it out

A viral X thread from Ben Cera claimed that Polsia raised $30 million at a $250 million valuation while approaching a $10 million annual run rate, with one founder and no employees.

The core pitch is bold: Polsia is an orchestration layer of agents that can help run company functions end to end, including investor replies, data-room management, product work, outbound, and operations. So much so that I am experimenting with it and so far it has been working well. I would say it is more like having a structured employee than what I experienced with agents like OpenClaw and Hermes.

The thread drew attention because it frames AI agents not as employee assistants, but as the operating system for a tiny company. It also attracted skepticism in the replies, including complaints about product quality, outbound spam, and whether the results match the fundraising story.

That tension is the real story: investors are funding the possibility of companies with radically smaller teams, while customers and operators are still testing whether the agent stack can carry real business responsibility.

SO WHAT?

Tiny AI-native teams are becoming investable, but credibility will depend on retention, customer outcomes, and whether the automation creates leverage instead of noise.

See the viral post on X:

A popular X post showed how people are using image models to turn personal photos into minimalist double-exposure travel posters with one prompt.

The shared prompt asks the model to place a recognizable traveler silhouette inside a clean destination-themed design, mixing personal identity with a polished poster aesthetic.

For brands, travel creators, and local tourism businesses, this kind of lightweight generation can turn user photos into campaign assets, social posts, and personalized merch concepts almost instantly.

Get the prompt on X:

The next wave of millionaires will be people who figured out how to make AI work for them.

The window to get ahead is still open. But not for long.

Here are 200+ proven ways to make money with AI in 2026.

Sign up for Superhuman AI, the free daily newsletter read by 1M+ professionals, and get instant access to all 200+ ways to profit from AI this year.

Claim your free list

Apple's long-delayed Siri overhaul may be taking a more open path. A 9to5Mac report, citing Bloomberg coverage, says future Apple AI features could let users choose third-party models from companies such as Google and Anthropic for Siri, Writing Tools, and related experiences.

Earlier this year, Apple had already confirmed that Google's Gemini models would help power future Siri capabilities.

The shift matters because Apple no longer has to rely only on its own model progress to make Siri useful. If the company allows outside models to plug into the system, users could bring their preferred AI assistant into Apple's interface while Apple keeps control of the device experience.

The big unresolved question is execution: privacy, app integration, voice experience, and whether Apple can turn model choice into a simple consumer feature rather than another settings menu.

SO WHAT?

If Apple opens Siri to leading AI models, the iPhone could become a distribution layer for AI assistants rather than a closed bet on one in-house model. So far, Siri has not been impressive, will this turn the tide for Apple?

Read the source:

Google DeepMind CEO Demis Hassabis shortened his public AGI timeline, saying the technology could arrive around 2029 or 2030.

Sherwood reported that Hassabis had previously pointed to a wider 2030 to 2035 window, but after Google I/O he argued that rapid progress in agents makes an earlier date easier to imagine.

The article places Hassabis alongside a growing list of AI leaders who now talk about AGI in near-term planning windows rather than distant science fiction.

The predictions still vary widely, and definitions of AGI remain slippery, but the direction is clear: the people building the frontier systems increasingly expect more autonomous, broadly capable agents before the end of the decade.

SO WHAT?

Whether or not 2030 is right, leaders should treat agentic AI as a strategic planning issue now, not as a future trend to revisit later. If your company is not ready for Agentic AI, then forget about planning for AGI!

Read the source: