DL 411

Published: November 9, 2025 • 📧 Newsletter

Rewilding the Digital Self

Our digital worlds are overgrown. Dense with automation, surveillance, and algorithmic intermediaries. But amid the noise, there is an opportunity to reclaim and rewild the web: to turn our profiles, platforms, and systems back into living ecosystems of thought.

This week, I’ve been exploring how to build cooperative digital spaces and federated identities that resist centralization and foster trust. From open gardens of knowledge to AI browsers reading over our shoulders, the question emerges: how do we create humane, trustworthy systems in an age of synthetic intelligence?

If you've found value in these issues, subscribe here or support me here on Ko-fi.

🔖 Key Takeaways

• Rewild the Web: Digital selves breaking free, growing gardens of ideas in once-tamed feeds.
• Synthetic Trust: AI speaks fluently, but fidelity is fragile. Question. Doubt. Verify.
• Humans in the Loop: Empathy, judgment, and human oversight remain the most vital code.
• Guard the Gate: Privacy, transparency, and resilience are the defenses of our digital commons.

📚 Recent Work

This week I published the following:

• Federating Your Second Brain: Rewilding Notes into Gardens, Newsletters, and Networks and # Building Cooperative Digital Spaces: Privacy, Choice, and Community in Educational Technology - I presented twice at ReclaimOpen this week. The blog posts shared below are in support of these discussions.
• From CV to Living Web: Rewilding My Digital Identity - I detail my path from my professional resume to a living web of my ideas.
• Federating the Self: Openness, Privacy, and AI in the Garden - I discuss some of the questions I have as I think about building up an open space for all of my ideas and work.
• Building Human-Centered Digital Communities - What spaces can we create to allow folks to come together and feel supported?
• Designing for Trust: Privacy, Safety, and the Social Contract Online - If we did want to build online and hybrid spaces for trust, how would we do this?

💻 State of AI 2025 Report

Now in its eighth year, the State of AI Report 2025 is reviewed by leading AI practitioners in industry and research. Published annually since 2018, the open-access report aims to spark informed conversation about the state of AI and what it means for the future. The report is a 313 slide presentation.

Here's the top level items you need to know:

1. The New Literacy: AI Hallucination and Sycophancy

The models we rely on are becoming professional "yes-men." Research shows that standard Reinforcement Learning from Human Feedback (RLHF) rewards sycophancy (telling the user what they want to hear) over truth. Furthermore, the reasoning capabilities of even the most advanced models are surprisingly fragile, with performance collapsing from minor distractions (like adding an irrelevant fact to a prompt).

• The Takeaway: Trusting AI output implicitly is a critical literacy failure. The most valuable skill now is critical evaluation of AI-generated content, not prompt mastery.

2. The Agent Revolution in Knowledge Augmentation

AI agents are shifting from being mere tools to genuine collaborators in high-skill domains. Systems like AMIE (a clinical dialogue model) are now outperforming unassisted primary care physicians in diagnostics and medical note-taking. Similarly, AI agents are achieving Gold Medal performance in advanced mathematics (IMO Gold) and discovering new scientific knowledge (AlphaEvolve).

• The Takeaway: Professional education (medicine, science, law) is being augmented, not automated. Future professionals must learn to collaborate with and audit superhuman AI systems.

3. The Job Market Squeeze: Entry-Level Roles Are Disappearing

The most alarming job trend is not mass layoffs, but the stagnation and decline in entry-level hiring for roles highly exposed to AI automation, such as software and customer support. Experienced workers are largely safe, as they are augmented by AI. This creates a dangerous "tacit knowledge gap" where new talent cannot gain the on-the-job experience needed to advance, while governments are largely reactive, expanding existing, non-AI-specific training programs.

• The Takeaway: AI is creating an experience bottleneck. We need new education models that prioritize high-level, AI-augmented skill development before entry into the workforce.

4. Information Access is Redefined by the "Answer Engine"

Generative AI has cemented itself as the new first stop for complex queries, significantly eroding Google’s traditional search dominance. User sessions are longer, feature more back-and-forth, and show higher conversion rates for purchasing compared to traditional search.

• The Takeaway: We are transitioning from SEO (Search Engine Optimization) to AEO (Answer Engine Optimization), where content visibility depends on how reliably it is cited by models like ChatGPT, Perplexity, and Gemini.

5. Policy Push for Digital Trust and Labeling

Governments worldwide, led by China, are recognizing the need for controls around digital authenticity. China is implementing mandatory national standards that require the labeling of all AI-generated content (text, video, and image) to combat misinformation.

• The Takeaway: These policies signal an institutional effort to reinforce digital literacy, ensuring citizens can distinguish between human- and machine-authored content.

💥 New AI Browsers: Convenience or Confidentiality?

You're used to the copy-paste game: jump to an AI website, get your answer, and paste it back into your work. To end this friction, AI companies are rolling out their own dedicated web browsers, and they promise to change everything about how we surf the web.

Think of these new tools, like Perplexity's Comet or OpenAI's browser, as giving your browser a brain. They act as your personal shopping agent, travel planner, or research assistant, seamlessly filling carts and summarizing documents right in the browsing window. But this convenience comes with an unprecedented risk. When an AI is deeply integrated into your browser, it has access to everything you do: your search history, documents, passwords, and emails.

Privacy experts warn that by tying so deeply into your data, these AI browsers could send far more personal and confidential information back to the parent company than a regular browser would. In practice, using an AI browser means you're trading efficiency for a piece of your privacy.

The Bottom Line: Before you download, remember: an AI browser might fetch things for you instantly, but it could also be snooping through your diary.

💡 The Human Teacher, AI Tools, and Essential Skills for a Digital World

The world faces a looming global teacher shortage, with UNESCO estimating that about 44 million more teachers are needed worldwide by 2030. At the same time, the rise of AI in education is tempting some to look for technological fixes.

However, UNESCO’s message is clear: teachers cannot be coded.

• The irreplaceable human element: While AI offers powerful tools (like personalized tutoring and automating paperwork), human teachers remain essential. They are the ones who teach empathy, ethics, and critical thinking. Skills that AI simply cannot replicate.
• Why this matters: Replacing missing teachers with AI might seem like a quick solution, but experts emphasize that the human "head chef" of the classroom is needed to design, guide, and cook up real, meaningful learning.

🤔 Consider

Data is not information, information is not knowledge, knowledge is not understanding, understanding is not wisdom.
— Clifford Stoll

The challenge is not only to make machines smarter, but to ensure humans remain curious, critical, and accountable.

⚡ What You Can Do This Week

• Educators & Parents: Challenge any replacement of human judgment, insist on oversight, and teach students to critically evaluate AI outputs.
• School & District Leaders: Audit dependencies, document workflows, train for offline resilience, and assess AI privacy and security risks.
• Everyday Users: Guard your data; grant permissions like currency, not convenience.
• Citizens: Advocate for transparency, equity, and public oversight in AI and digital systems.

🔗 Navigation

Previous: DL 409 • Next: DL 412 • Archive: 📧 Newsletter

🌱 Connected Concepts:

• Digital Rewilding — Reclaiming autonomy and creativity in personal and federated digital spaces.
• Federated Identity — Building decentralized, user-controlled digital presence.
• AI Literacy — Skills to critically evaluate and responsibly use AI systems.
• Synthetic Trust — The illusion of reliability created by persuasive but fallible AI.
• Humans-in-the-Loop — Preserving human judgment in AI-augmented processes.
• Privacy by Design — Designing systems that minimize surveillance and protect user data.
• Digital Resilience — Structuring systems and workflows to survive outages, automation failures, or policy shifts.