Notes on thinking in the AI era.
Source-backed editorial on cognitive debt, AI brain fry, skill atrophy, daily brain exercise habits, and the research behind why deliberate practice still matters.
- ai news
The Anthropic coding-skill study: what developers should actually take away
Anthropic's 2026 study found AI-assisted developers scored ~17 points lower (50% vs 67%) on an immediate comprehension quiz when learning a new library. A careful read of what it does — and doesn't — show, and what developers should change in their daily practice.
- for lawyers
Brain exercise for lawyers and paralegals in the AI era
Legal-tech tools shift drafting and research onto AI, but the close-reading and judgment muscles still belong to the lawyer. A practice routine for the working day.
- how to
Brain-exercise routines: five common formats, honestly compared
Single-app daily, multi-app stack, book-based, structured course, no-app practice. The five common formats of a daily thinking habit, compared honestly — without claiming any one is best.
- category explainers
Brain exercise vs brain training — a 2026 explainer
Same surface activity, different promise structure. The category distinction matters because the FTC sanctioned the broader of the two promises in 2016, and the 2014 Stanford consensus targeted the same marketing claims.
- students and learning
ChatGPT for studying — when it helps, when it hurts
An honest, sourced playbook for students using ChatGPT. The published research on AI-and-learning outcomes, where AI genuinely helps, and where it produces the artefact-good thinking-weaker pattern that researchers have documented.
- how to
Daily code reading without autocomplete: a six-rule playbook for developers
AI assistants are excellent at writing code. They are not the same as reading code. A practical six-rule playbook for keeping the code-reading skill alive in the AI-coding era.
- for over 50s
Daily mental practice after 50 — an honest guide
What the published evidence actually supports for cognitive health after 50, what the FTC took action against, and a daily practice routine that does not overclaim.
- for writers
Daily practice for writers in the AI era
What working writers can do to keep their own voice while AI tools take over more of the drafting pipeline. Cited evidence plus a concrete practice routine.
- skills deep dive
Daily writing practice when AI writes faster
A deep-dive into what daily writing practice actually looks like in 2026, why the originating act matters, and the published evidence behind the calibration position.
- skills deep dive
Deep reading when everything summarises
What daily deep-reading practice looks like in 2026, the published evidence on AI summarisation and critical thinking, and why sustained attention is a habit that responds to use.
- for designers
Design judgment and AI generative tools
Figma AI, Midjourney, and the rest of the generative stack produce options at scale. The cognitive act that defines a designer is deciding — not generating. A practice routine for designers in 2026.
- research explainers
Do brain training apps actually work? A 2026 review of the evidence
What the published research — Stanford 2014 consensus, Simons 2016 review, NIH ACTIVE Trial, FTC v Lumosity — actually says about whether commercial brain-training apps work, and what changed in 2026.
- for executives
Executive judgment when everything is AI-mediated
Leadership judgment in a working day that runs through AI from end to end. What the 2026 research suggests, the team-level modelling effect, and a defensible practice routine.
- how to
Five daily brain-exercise habits that actually stick
Habit-design rules grounded in BJ Fogg's Tiny Habits and Ericsson's deliberate-practice research. Five rules for installing a daily thinking practice that survives a busy month.
- for founders
Founder judgment in the AI era
The originating cognitive acts of running a company — judgment under uncertainty, deciding before knowing — are exactly the ones AI most easily takes over. A practice routine for shipping-heavy schedules.
- category explainers
Free vs paid brain exercise apps in 2026 — what you actually get
An honest, sourced look at the free-vs-paid question in the brain-exercise category. What the ad-supported tier actually costs you, what subscriptions tend to unlock, and which trade-off makes sense.
- research explainers
GPS and navigation skill — what actually changed
The cleanest empirical case for cognitive offloading. UCL's 2020 study, the UCSB extension, the mechanism, and what it means for AI tools by analogy.
- research explainers
GPS changed our memory — what AI might do too
The 2020 UCL study is the cleanest natural experiment in the cognitive-offloading literature. The mechanism is direct, the parallel to AI is real, and the practical advice is the same.
- category explainers
How much brain exercise per day is enough?
An honest, sourced answer to the daily-dose question. The published deliberate-practice and habit-design literature points to a smaller number than most marketing implies — 5 to 15 minutes a day is enough.
- ai and thinking
Is ChatGPT making you dumber? An honest 2026 answer
What the published evidence actually supports — the MIT cognitive-debt preprint, the Stanković critique, the broader cognitive-offloading literature, and the honest framing the headlines miss.
- skills deep dive
Memory practice for the AI era — active recall
What active recall is, why it matters more in the AI era than before, and a daily practice routine grounded in the Google-effect literature and the GPS-and-spatial-memory studies.
- skills deep dive
Mental math in the calculator and AI era
What daily mental-math practice looks like when every device has a calculator and every browser has an AI. The published evidence on cognitive offloading and the case for keeping number sense warm.
- research explainers
Mental sharpness with age: the honest picture
What the cognitive-aging research actually says about fluid versus crystallized cognition, the lifestyle factors with real evidence, and why brain-training-prevents-decline claims drew regulatory action.
- how to
A morning thinking stack: Wordle, Senwitt, and a deliberate caffeine break
A practical morning ritual built on habit-stacking research. Stack Wordle (or the NYT Mini), Senwitt's seven-minute Set, and a deliberate caffeine break to install a daily thinking habit that survives any week.
- ai news
OpenAI's reasoning models: do they fix the thinking problem?
OpenAI's o-series reasoning models 'think before answering' using chain-of-thought-style processing. That changes what the model does. It does not change what the user does — and the cognitive-offloading concern is about the user.
- skills deep dive
Reasoning practice when ChatGPT thinks for you
What daily reasoning practice looks like in 2026, the published evidence on AI mediation and critical thinking, and a practice routine that keeps the originating cognitive act on the calendar.
- how to
Seven ways to keep writing sharp while using ChatGPT
A practical seven-item playbook for writers who use ChatGPT daily and want to keep their own voice. Grounded in the published cognitive-debt and cognitive-offloading research.
- how to
How to stop relying on AI for writing — a working playbook
A practical playbook for writers who want to keep their own voice while still using AI productively. Six concrete rules grounded in the published cognitive-debt research.
- for students
Student brain habits when AI does the homework
What the published evidence says about AI in student learning, the artefact-and-thinking paradox, and a study-habits playbook that keeps the originating cognitive acts on the timetable.
- for teachers
Teaching deliberate practice in an AI classroom
What teachers can do when students bring AI to every assignment. Modelling the originating cognitive act, redesigning assessment, and keeping deliberate practice in the room.
- ai news
Ten years after FTC v Lumosity: what actually changed in brain apps
The 2016 FTC settlement against Lumosity turns ten. The honest accounting of what the category learned, what it didn't, and what the post-2016 honest products look like.
- ai news
The BCG/HBR 'AI brain fry' study, explained
A focused study-explainer on the 2026 BCG and HBR-aligned research covered as 'AI brain fry'. What the ~14% figure does and doesn't show, why the mechanism is plausible, and what's missing from most coverage.
- research explainers
The Google effect and what it means for AI
The 2011 Sparrow et al. paper started the cognitive-offloading conversation in earnest. The 2024 meta-analysis sharpened the picture. The AI extension is the natural next step.
- research explainers
Transactive memory: when AI is the partner
Daniel Wegner's 1985 framework explained how long-term couples share a single memory system. In 2011, Sparrow et al. extended the frame to search engines. AI is the next layer.
- research explainers
Use it or lose it: the skill-decay literature
The Arthur et al. 1998 meta-analysis is the canonical reference for skill decay. The deliberate-practice tradition explains the inverse. The AI moment is the reason the pattern matters now.
- ai news
What Senwitt borrows from Duolingo — and what it doesn't
Duolingo's habit design and streak research are the most documented in consumer learning. Senwitt borrows the bounded daily ask and the forgiveness on missed days. We don't borrow the streak guilt or the leaderboard pressure.
- research explainers
What the research actually supports about brain exercise
The exact claim line, with citations: what the cognitive literature supports about practising thinking skills, what it doesn't, and why Senwitt's marketing language sits where it does.
- category explainers
When should you start brain exercise — at 30, 50, or 70?
An honest, sourced answer to the 'what age should I start' question — what cognitive aging research actually supports, what it doesn't, and why the timing question is less load-bearing than the daily-habit question.
- compare
Brain training app comparison — 11 apps, 2026, no overclaim
A blog-length comparison of 11 apps in and around the brain-exercise / brain-training / daily-puzzle category — BrainHQ, Lumosity, Elevate, Peak, NeuroNation, CogniFit, memoryOS, NYT Games, Wordle, Duolingo, and Senwitt. The deeper version is in the guide.
- ai and thinking
How to use ChatGPT without losing your edge
The MIT cognitive-debt paper documented what daily ChatGPT use does to the cognitive surface underneath your work. Four habits that consistently work — and the daily-practice block that ties them together.
- habit
The morning ritual for knowledge workers in the AI era
AI changed what mornings cost knowledge workers. The old pre-work block — coffee, reading, the first deliberate thinking of the day — is now competing with the ChatGPT tab. A 2026 morning ritual that survives.
- product philosophy
Why seven minutes — the case for the smallest daily practice block
The seven-minute daily Set isn't arbitrary. It's the smallest block that fits a meaningful mixed-rep session across multiple Skills while staying short enough to survive any day. Here's the design reasoning.
- product philosophy
Why Senwitt is built like Peloton (for the brain)
Peloton solved daily-habit fitness with streaks, instructors, personal-bests, and a defined practice surface. Senwitt borrows the structural pattern for daily-habit brain exercise — and deliberately leaves out the parts that don't translate.
- compare
Wordle and NYT Games — an honest look at the daily-puzzle category
Wordle and NYT Games defined the modern daily-puzzle ritual. They are not brain exercise apps, but they share the daily-habit slot. Here is what they actually train, where they fit, and where Senwitt is a different job.
- ai and thinking
An honest self-assessment for AI dependency (no clinical pretense)
A self-check based on Talkspace, Psychology Today, and HBR coverage of AI dependency in knowledge workers — symptoms, signals, and a small habit response that doesn't require quitting AI.
- ai and thinking
Is ChatGPT actually making students lazy? Sorting the evidence
A close read of the MIT cognitive debt study, Cal Newport's New Yorker essay, and The Conversation's earlier coverage — and what the evidence actually supports about ChatGPT and student thinking.
- habits
A five-minute daily thinking habit, designed for AI-heavy days
A practical five-to-seven-minute daily routine for keeping writing, math, code, memory, reading, and reasoning in regular practice when the rest of your day runs on AI.
- ai and thinking
Avoiding skill atrophy: how engineers can use AI coding tools without losing the craft
Anthropic's 2026 study measured a 17% drop in developer skill formation when AI coding assistants did the heavy lifting. Here is how to use AI without paying that cost.
- research
What the MIT cognitive debt study actually shows (and what reporters keep getting wrong)
A careful read of the 2025 MIT Media Lab cognitive debt study — what it tested, what it found, what coverage often overstates, and what it means for daily practice.
- ai and thinking
AI brain fry: why your head feels foggy after a day on AI chatbots
A cited explainer on 'AI brain fry' — the cognitive fatigue knowledge workers report after long stretches of AI use, what BCG and HBR have measured, and how to keep practice in your day.