AI Is Changing Books, Work and Software: Ten Topics from Tim Ferriss and Forward Future

Chatbots are putting pressure on how-to books, and the big AI companies are competing to become the main work assistant. This overview sums up ten topics from Tim Ferriss’s blog and podcast and from Forward Future, plus a product launch from OpenAI.

1. Has AI already “killed” how-to books?

In a blog post from June 12, 2026, Tim Ferriss uses his own sales figures (BookScan) to show how sharply his how-to books are declining: his print catalog fell about −46% in 2025 compared with the previous year, and 2026 is running at about −57% compared with 2025. His explanation: a book used to be the best “interface” for getting answers (for example, losing weight or sleeping better). Today millions of people prefer to ask a free chatbot (a conversational AI such as ChatGPT or Claude), which produces a personal plan in seconds.

He does not mean that books are dying entirely. What is under pressure is mainly pure instructional information (“the 5 steps to X”). What is more likely to last: real stories, personality and content that genuinely changes people. His advice: write for 10,000 true fans rather than for millions of short videos that are forgotten right away.

Original article: https://tim.blog/2026/06/12/has-ai-already-killed-nonfiction/

2. “Better than free”: standing out in the AI era

On July 17, 2026, Ferriss published an updated essay by Kevin Kelly (co-founder of WIRED). The core idea: when AI produces copies and variations of text, images, music and advice almost for free, you have to offer things that cannot be copied. Kelly names eight such values (“generatives”): immediacy, personalization, interpretation (help with understanding), authenticity, accessibility, embodiment (for example, a live performance), patronage (supporting the creators) and findability.

Trust, relationships and context therefore become more valuable than “yet another AI text”. Examples mentioned include Brandon Sanderson, Stratechery and Daily Stoic, where fans pay for experiences, interpretation or premium editions.

Original article: https://tim.blog/2026/07/17/better-than-free/

3. Podcast: Fei-Fei Li on spatial intelligence (#839)

Episode #839 (December 9, 2025) features Dr. Fei-Fei Li: Stanford professor, co-founder of World Labs, often called the “Godmother of AI”. Among other things, she was involved in ImageNet (a large image database that greatly helped modern image recognition) and describes AI as a civilizational technology that changes the economy, culture and everyday life.

Her current focus is spatial intelligence: AI that understands space, 3D and how things relate in the physical world, beyond plain text. For her, the people who build, use and are affected by AI are at the center.

Original article: https://tim.blog/2025/12/09/dr-fei-fei-li-the-godmother-of-ai/

4. Podcast: Q&A on the “AI tsunami” (#859)

In Q&A episode #859 (March 26, 2026), Ferriss answers listener questions about the coming wave of AI. He says that with technology he often chooses the “dull edge” rather than the “bleeding edge”, meaning proven, stable tools rather than every newest experiment. Topics: what humans still do better than AI, what AI should not be used for, how to stay visible in the flood of AI content, and how his team uses AI internally.

Other topics include career changes at a time when AI is changing jobs, Kevin Kelly’s idea of 1,000 true fans, and the “offline advantage”: relationships and presence that no chat can replace.

Original article: https://tim.blog/2026/03/26/qa-with-tim-ai-tsunami/

5. Podcast: Sebastian Mallaby on the race to superintelligence (#870)

Episode #870 (June 17, 2026) with author Sebastian Mallaby about his book The Infinity Machine (about Demis Hassabis, DeepMind and the pursuit of superintelligence, meaning AI that surpasses humans in many areas). Mallaby interviewed more than 100 insiders. Key points: you should be excited and cautious at the same time; the probability of “worst outcomes” is not zero; and one lab head expects that by around 2028 recursive self-improvement (models that help develop the next model themselves) could greatly accelerate the race.

This involves big opportunities and real risks such as deception, loss of control and the geopolitical rivalry between the US and China.

Original article: https://tim.blog/2026/06/17/sebastian-mallaby/

6. Why every major AI company wants to be the assistant

An article by Nick Wentz (Forward Future, October 9, 2026): it used to be mainly about which model is the smartest. Now OpenAI, Google, Microsoft, Anthropic, Meta and Grok Bot are competing to be the work assistant, the place where the workday begins. The reason is context (files, colleagues, preferences, history): it makes an assistant hard to replace. This is called lock-in (being tied to one provider).

Instead of opening email, cloud storage and a CRM (customer relationship management software) separately, you give the assistant a task. The software underneath stays important, but the “front door” to work is moving. Wentz compares six strategies: OpenAI as an independent layer, Google with Workspace context, Meta with everyday agents, Microsoft for large companies, Anthropic for power users, and Grok Bot as “just give it the job”.

Original article: https://www.forwardfuture.ai/newsletter/originals/why-every-ai-company-wants-to-be-your-assistant

7. The new power user manages AI workers

Also by Nick Wentz, October 4, 2026: the big productivity gap is no longer just between “uses AI” and “doesn’t use AI”, but between people who query AI like a tool and those who manage teams of AI agents (programs that complete multi-step tasks on their own).

Four skills: delegating (describing the outcome clearly), planning handoffs between agents, defining quality in advance, and improving the system when errors occur instead of just fixing a single result. The goal is management by exception: routine work runs on its own, and people step in only for exceptions. As a starting point, the article recommends setting up a recurring process (for example, competitor research) so that you remove yourself from most steps.

Original article: https://www.forwardfuture.ai/newsletter/originals/the-new-power-user-is-a-manager-of-ai-workers

8. Six takeaways from OpenAI DevDay

Alex Salkever (October 2, 2026) on OpenAI’s developer conference: agents are at the center. OpenAI mentions, among other things, “Dots” (personal agents with their own cloud environment, GPT-6 Astra and many plugins, i.e. extensions) and cites around 1.2 billion weekly users as its reach.

Other points: startups discover new needs and big providers follow very quickly; software companies (Salesforce, Adobe, Figma …) are integrated through a marketplace and plugins rather than replaced; users will soon expect near-real-time answers; connecting multiple email accounts makes agents “personal”; and mobile devices become more of a remote control for agents working elsewhere.

Original article: https://www.forwardfuture.ai/newsletter/originals/six-takeaways-from-openai-s-big-devday

9. Vibe CAD: AI designs circuit boards with GPT-6 “Astra”

Aleksandar Kostovic, October 1, 2026: “vibe coding” meant having software built from everyday language. Now comes “vibe CAD”: the model works inside CAD programs (software for technical drawing and design, e.g. KiCad), places components, routes circuit boards (PCBs, printed circuit boards) and applies design rules.

Community examples: a Bluetooth wearable in about 30 minutes and an FPV drone board that was actually manufactured. Humans remain important for specification, review and production readiness; the AI increasingly takes over execution in CAD.

Original article: https://www.forwardfuture.ai/newsletter/originals/vibe-cad-engineering-with-gpt-6-astra

10. What people actually do with OpenClaw: 25+ use cases

A guide from Forward Future (as of early February 2026): OpenClaw is an open-source agent system (freely available source code) that lets users run automations through conversation: email, CRM, briefings, development, content, smart home. It collects real deployments such as morning briefings, video pipelines, coding from a phone and dashboards with multiple agents.

The guide is older than the October articles but shows what already works today. Recommended start: begin small and then expand.

Original article: https://www.forwardfuture.ai/p/what-people-are-actually-doing-with-openclaw-25-use-cases

Bonus: GPT-6 with “Intelligent UI” for almost all ChatGPT users

On October 7, 2026, OpenAI announced that, after paying customers, a GPT-6 version is rolling out to more than 1.2 billion weekly ChatGPT users. New: answers can be interactive interfaces, such as charts, buttons, forms and small tools right in the chat.

In addition, answers can start while the model is still “thinking”; with web search, according to OpenAI, answers start earlier on average. Rollout order: Plus/Pro/Business/Enterprise first, then Free/Go.

Original article: https://openai.com/index/gpt-6-for-everyone/