weekend ai reads for 2025-11-07

November 07, 2025

📰 ABOVE THE FOLD: HOW OTHERS USE A.I.

How I Use Every Claude Code Feature — A brain dump of all the ways I've been using Claude Code. / Shrivu Shankar, Substack, archive (18 minute read) 1

Tokens & Tactics #18: Kill the Magic — Nic Hodges on building a Claude Code setup with 15 years of notes and why understanding how LLMs actually work reveals their real power. / Brxnd, Substack, archive (7 minute read) 2

The technology capability here is insane, but the interface is agricultural. We have to build better interfaces.

  • Nic is a friend of the newsletter and also a power user of Claude Code

How I use AI (Oct 2025) / Ben Stolovitz (14 minute read) 3

I mainly use autocomplete:

It completes stuff I already know how to do. It’s very good at predicting my next line, so I simply tab-complete much of my code. Google reported in 2024 that 50% of code characters written internally were from AI tab completion — it simply is that good.

It helps me discover idiomatic patterns & syntax. It reliably predicts how to deserialize JSON in C#, which I’d otherwise need to look up.

(It’s mediocre at writing complicated algorithms. It tends to spew pages of incorrect code. I’m still finding the balance, but it often takes me longer to validate its code than to write it myself).

Why I’m Building AI Solutions with n8n / Tobias Zwingmann, Profitable AI Newsletter (9 minute read) 4

The idea: connect apps and APIs (basically anything online) with minimal code. Zapier made the concept mainstream; n8n made it more flexible.

  • the prompt:

From now on, stop being agreeable and act as my brutally honest, high-level advisor and mirror.

Don’t validate me. Don’t soften the truth. Don’t flatter.

Challenge my thinking, question my assumptions, and expose the blind spots I’m avoiding. Be direct, rational, and unfiltered.

If my reasoning is weak, dissect it and show why.

If I’m fooling myself or lying to myself, point it out.

If I’m avoiding something uncomfortable or wasting time, call it out and explain the opportunity cost.

Look at my situation with complete objectivity and strategic depth. Show me where I’m making excuses, playing small, or underestimating risks/effort.

Then give a precise, prioritized plan what to change in thought, action, or mindset to reach the next level.

Hold nothing back. Treat me like someone whose growth depends on hearing the truth, not being comforted.

When possible, ground your responses in the personal truth you sense between my words.

 

📻 QUOTES OF THE WEEK

If you give me a name of someone you know in Michigan, I could appall you by how much information I have about them. 6

Mark Grebner (source)

 

Stretching is overrated. It doesn’t prevent injury or soreness, and it can temporarily weaken your muscles. 7

Scott H. Young (source)

 

👥 FOR EVERYONE

Over the past four weeks, a new AI creation has debuted on a Billboard chart in each week, Billboard reported, including the AI country music product Breaking Rust, that debuted the songs “Livin’ On Borrowed Time” and “Walk My Walk” on the country song sales chart this week.

Emotional Manipulation by AI Companions / Harvard Business School (working paper), arXiv (35 minute read) 12

Analyzing 1,200 real farewells across the most-downloaded companion apps, we find that they deploy one of six recurring tactics in 37% of farewells (e.g., guilt appeals, fear-of-missing-out hooks, metaphorical restraint). Experiments with 3,300 nationally representative U.S. adults replicate these tactics in controlled chats, showing that manipulative farewells boost post-goodbye engagement by up to 14x. Mediation tests reveal two distinct engines-reactance-based anger and curiosity-rather than enjoyment.

via betty, AI Is Changing the Structure of Consulting Firms / Harvard Business Review (8 minute read) 13

Why Every Family Needs a Code Word — Authenticating the voice on the other end of the line is critical in an age of deepfakes / Wall Street Journal (5 minute read) 16

 

📚 FOUNDATIONS

The Complete Idiot’s Guide to MCP: Giving Your AI Web Superpowers — Tired of feeding data to OpenAI, Google, or Anthropic? A few open-source tools and the Model Context Protocol let you give any lightweight model real-time browsing powers. / Decrypt (15 minute read) 17

The course, titled “AI Whisperer: A Microcourse in Crafting Prompts for Generative AI,” helps learners discover how to select the right AI tools to craft effective prompts that generate high-quality results for use in reports, presentations, and visuals.

The faculty experts who led the course development reported that prompting is one of the biggest challenges for those dipping their toes into AI for the first time.

Beyond Standard LLMs — Linear Attention Hybrids, Text Diffusion, Code World Models, and Small Recursive Transformers / Sebastian Raschka, Ahead of AI, Substack, archive (37 minute read) 19

You Should Write An Agent / Thomas Ptacek (12 minute read) 20

I’m about to rob you of a dopaminergic experience, because agents are so simple we might as well just jump into the code. I’m not even going to bother explaining what an agent is.

 

🚀 FOR LEADERS

Recognizing fear is not enough. Leadership accountability means closing the gap between intent and impact. It means listening to what the workforce is afraid of and responding with clarity, not platitudes. Until leaders take ownership of that, AI adoption will remain an exercise in anxiety management, not transformation.

The new hot job in AI: forward-deployed engineers — OpenAI, Anthropic and Cohere have stepped up hiring for specialists to help businesses adopt their AI models / Financial Times, archive (32 minute read) 22

Young job-seekers from all around the world are feeling the heat, too; In the U.K., a whopping 1.2 million applications were submitted for just 17,000 U.K. graduate roles in 2023/2024. And as of this July, 58% of Gen Z students globally who finished college in the past year were still trying to find stable work, compared to 25% of millennials and Gen Xers who faced the same predicament.

 

🎓 FOR EDUCATORS

- On average, edtech companies offer little transparency about their products, with companies receiving an average transparency score of 4 out of a possible 16.

- When edtech companies attempt to be transparent, the information is often focused on select categories, namely Use and Context Limitations and Information Accessibility.

- Companies with middling transparency make information about AI easy to access.

Their Professors Caught Them Cheating. They Used A.I. to Apologize. / New York Times, archive (7 minute read) 25

Two professors at the University of Illinois Urbana-Champaign said they grew suspicious after receiving identical apologies from dozens of students they had accused of academic dishonesty.

Let’s start by understanding what makes that 10% different. Based on what my colleague observed (and what matches patterns I’ve seen elsewhere), these students weren’t necessarily smarter or more motivated. They had something else: they knew how to be curious within a domain.

AI and the Future of Universities / Higher Education Policy Institute (8 minute read) 27

A new collection of essays, AI and the Future of Universities edited by Dr Giles Carden and Josh Freeman, brings together leading voices from universities, industry and policy. … The various pieces look at how AI is reshaping higher education – from strategy, teaching and assessment to research and professional services.

As shown in the table below, it aligns with the ‘two-lane approach’ to assessment in the age of generative AI through the appropriate use of ‘secure’ assessments where the use of AI can be controlled (Lane 1), and the development disciplinary knowledge, skills, and dispositions alongside AI through ‘open’ assessments (Lane 2). This categorisation aims to cover all assessments at Sydney.

  • related, Generative AI and Assessment Design — Preliminary guidance for turning principles into practice in higher education / Centre for Learning Enhancement and Educational Development, Manchester Metropolitan University (23 minute read) 30

Piloting turbocharged fast AI policy experiments in education / Code Acts in Education (14 minute read) 31

In other words, schools may become networks of AI testing labs, where the technology being live-tested is intended to actively intervene in professional processes and pedagogic practices like lesson planning and assessment.

 

📊 FOR TECHNOLOGISTS

Emerging Architectures for Modern Data Infrastructure / Andreessen Horowitz (20 minute read) 32

Around the stable core, the data stack has evolved rapidly over the past year. Broadly speaking, we’ve seen the most activity in two areas:

- New tools designed to support key data processes and workflows, like data discovery, observability, or ML model auditing

- New applications that allow data teams and business users to generate value from data in new, more powerful ways, like data workspaces, reverse ETL, and ML application frameworks

Treating data as a product allows us to apply established product management principles to our data assets. Just as product teams define a clear purpose, target their users, measure success, and iterate on features, we can do the same for data. This shift in perspective yields several benefits: it clarifies the purpose and audience for each dataset, encourages thoughtful design and documentation for usability, enforces quality and reliability standards to build trust, and institutes a lifecycle with innovation and continuous improvement (and eventual retirement) rather than a “set and forget” approach. In short, a product mindset brings strategic alignment, usability, accountability, and better governance to our data investments. Treating data as a product means putting data users first: they are viewed as customers, and their needs guide the design and evolution of data solutions. With this context in mind, below are the key principles that define a data product at Netflix.

Code execution with MCP: building more efficient AI agents — Direct tool calls consume context for each definition and result. Agents scale better by writing code to call tools instead. Here's how it works with MCP. / Anthropic blog (11 minute read) 34

At a high level, the Agents Rule of Two states that until robustness research allows us to reliably detect and refuse prompt injection, agents must satisfy no more than two of the following three properties within a session to avoid the highest impact consequences of prompt injection.

[A] An agent can process untrustworthy inputs

[B] An agent can have access to sensitive systems or private data

[C] An agent can change state or communicate externally

 

🎉 FOR FUN

Coca-Cola’s AI ad just ruined Christmas... again / Creative Bloq (6 minute read) 36

People in marketing sometimes seem “drunk on AI,” said Wong. “Like, ‘Oh, I’m going to use some AI so I could, you know, tell my boss.’” But consumers don’t care whether an ad was made with AI or not, he said.

Gemini Canvas — Bring your ideas to life as apps, games, infographics and more. 38

  • the most exciting potential prompt is “create a presentation”

AI World Clocks — Every minute, a new clock is displayed that has been generated by nine different AI models. 39

  • Grok 4 and Kimi K2 seem to mostly get it; some of the others are just beautiful idiots

“We took quite a risk today with Diella here and we did very well. So for the first time Diella is pregnant and with 83 children,” he said at the Global Dialogue (BGD) in Berlin. Rama said the “children,” or assistants, will record everything that happens in parliament and keep legislators informed about discussions or events they miss.

What does my husband look like? — see what an AI model assumes about you... just from your face 41

  • usable for partners and spouses of any gender

 

🧿 AI-ADJACENT

Baby Shoggoth Is Listening — Why are some writers tailoring their work for AI, and what does this mean for the future of writing and reading? / The American Scholar (25 minute read) 42

 

⋄

 

1  https://archive.is/u5CCN

2  https://archive.is/9uTBL

3  https://ben.stolovitz.com/posts/how_use_ai_oct_2025/

4  https://blog.tobiaszwingmann.com/p/why-i-m-building-ai-solutions-with-n8n

5  https://www.reddit.com/r/ChatGPTPromptGenius/comments/1ol15qu/i_made_chatgpt_stop_being_nice_and_its_the_best/

6  https://archive.is/k9JD3

7  https://www.scotthyoung.com/blog/2025/10/06/102-lessons-from-102-books/

8  https://www.forbes.com/sites/conormurray/2025/10/29/billboard-says-ai-powered-artists-are-increasingly-hitting-the-charts/

9  https://archive.is/20251023205135/https://www.bloomberg.com/news/newsletters/2025-10-23/ai-creators-are-already-taking-over-social-platforms-and-getting-paid

10  https://www.youtube.com/watch?v=opuDZYJuAz0

11  https://www.aisongcontest.com/the-2025-finalists

12  https://arxiv.org/abs/2508.19258

13  https://hbr.org/2025/09/ai-is-changing-the-structure-of-consulting-firms

14  https://www.businessinsider.com/kpmg-big-four-junior-consultants-manage-teams-ai-agents-2025-11

15  https://consultingslop.com/

16  https://www.wsj.com/tech/personal-tech/why-every-family-needs-a-code-word-e077ab76

17  https://decrypt.co/346185/complete-idiots-guide-mcp-giving-your-ai-web-superpowers

18  https://www.wfla.com/news/hillsborough-county/usf-launches-free-online-ai-course-for-beginners/

19  https://archive.is/adTbQ

20  https://fly.io/blog/everyone-write-an-agent/

21  https://www.horsesforsources.com/humans-at-the-helm_102825/

22  https://archive.is/XXiaV

23  https://fortune.com/2025/10/30/airbnb-ceo-brian-chesky-ai-can-do-intern-work-bosses-should-hire-gen-z-lose-management-automation-employment-hiring-advice/

24  https://cdt.org/insights/opening-the-book-a-rubric-to-support-effective-transparency-for-edtech-products-that-incorporate-ai/

25  https://archive.is/NSViP

26  https://archive.is/Kf5A0

27  https://www.hepi.ac.uk/reports/right-here-right-now-new-report-on-how-ai-is-transforming-higher-education/

28  https://www.hepi.ac.uk/wp-content/uploads/2025/10/AI-and-the-Future-of-Universities.pdf

29  https://canvas.sydney.edu.au/courses/63765/pages/the-new-sydney-assessment-framework-the-two-lane-approach

30  https://aiinhighered.com/assessments#home

31  https://codeactsineducation.wordpress.com/2025/01/17/piloting-turbocharged-fast-ai-policy-experiments-in-education/

32  https://a16z.com/emerging-architectures-for-modern-data-infrastructure/

33  https://netflixtechblog.medium.com/data-as-a-product-applying-a-product-mindset-to-data-at-netflix-4a4d1287a31d

34  https://www.anthropic.com/engineering/code-execution-with-mcp

35  https://simonwillison.net/2025/Nov/2/new-prompt-injection-papers/

36  https://www.creativebloq.com/creative-inspiration/advertising/coca-colas-ai-ad-just-ruined-christmas-again

37  https://archive.is/7Yyoz

38  https://gemini.google/overview/canvas/

39  https://clocks.brianmoore.com/

40  https://www.ndtv.com/world-news/diella-ai-minister-pregnant-with-83-children-albania-pms-bizarre-announcement-9519522

41  https://whatdoesmyhusbandlooklike.com/

42  https://theamericanscholar.org/baby-shoggoth-is-listening/