Don’t Be Fred! 8 Bad Habits to Lose in 2025

You know that moment when someone tells a story so perfectly out-of-step with the times that you laugh—then realize you're guilty of your own version? That's exactly what happened when Brice Challamel, Head of AI at Moderna, shared the origin of their company mantra, "Don't be Fred" on Beyond the Prompt. Fred was a leader who literally hand-wrote email replies on printouts for his assistant to re-type. We all chuckled… until Brice pointed out the uncomfortable truth: in the age of AI, we're all at risk of being Fred. Behaviors that seemed normal just months ago now look like relics. Having caught myself being Fred more times than I care to admit, that insight really stuck with me (and honestly, left me with a bunch of follow-up questions). So I asked Brice to explore it further. Below, he digs into the real question: how do we recognize—and stop—our own Fred-like habits when better options are sitting right in front of us?

***

This morning, my 13 yo son paused his video game to browse YouTube for a character build. What were the chances he'd find the right video and watch it entirely in minutes?

My son is Internet-native, not AI-native. His reflexes lead him down an endless path designed for social media retention. I'm no different, and probably worse.

Despite working with AI for years, I still find myself scrolling emails, using sticky notes, or frantically opening folders to find documents.

It’s 2025. Why do we still do this to ourselves?

After a brief conversation, my son agreed to try another way. He used dictation to share his situation in ChatGPT and provided additional context, then let it browse a vast number of YouTube video transcripts and dozens of other sources. It came back in 1m 48s with a bespoke recommendation, including citations, optionality, and additional tips to make this build work for the battles ahead. And he offered to send me the link for this article.

As the Mandalorian would say, "This is the way!"

Choosing the right path in Hades. Hours by ourselves on the Web or minutes with AI?

No Shame, No Blame. We ALL Have Bad Habits...

The reason for this intro about my son is to get any shame linked to experience or age related to AI out of the way. The only humans born after the launch of GPT-4 are still learning to speak, so anyone reading this article would probably benefit from revisiting their habits, including myself and my teenage son.

This is not unprecedented. In the 1990s, a manager named Fred wouldn't use email. He had his assistant print his messages, hand-wrote his replies, and asked her to type them back into Outlook. It wasn’t simple resistance, it was the expression of a universal fatigue to engage with a new way of working, that delivers small comfort today at the cost of irrelevance tomorrow.

“Fred” is not a person of course, it’s a posture. Preferring familiar friction to a small up‑front investment in new capability. In 2025 and with the arrival of AI, the cost of being Fred is not cosmetic. It’s compounding: time lost to search, context lost to manual re‑typing of meeting notes, signal lost to busywork on sources add-up quickly.

But what does it mean concretely to be Fred nowadays, and what are the most rewarding small habits to identify and get rid of?

Growth Sometimes Means Letting Go of Something Precious

This article is not about the long list of super intelligent features advertised by SaaS incumbents that can memorize global legislation, triage support tickets or manage your expense claims.

It's not about spending dozens of hours Learning, watching tutorial videos, growing a second brain or programming swarms of bots and becomming super-human faster than the machine.

It’s about taking a step back and an honest look at the small, universal gestures we repeat every day: browsing, reading, drafting, listing, remembering. This is where Fred still hides and where we need to carve him out to reveal our better selves.

To make it practical, let’s frame each gesture with AI that these bad habits unlock not as a tool use but as a “Team of Five” collaboration:

• You: How to share context and clarify intent up-front.

• Assistant: what can be executed and fetched to remove friction.

• Expert: how a domain expert would sharpen or stress‑test it.

• Coach: how to extract feedback and learning for your future self.

• Creative Partner: how to widen or improve the idea.

Below are 8 Fred-like behaviors (each with an AI‑first alternative), and a simple score card to track them down and ruthlessly eliminate them from our lives.

1. Scroll through an email or chat thread

Fred‑like: Heroically catch up with your inbox every morning by scrolling messages. Open messages one by one, skim, try to reconstruct what changed, then paraphrase for yourself or others. Bonus points if you copy and paste manually the “important bits” into a new email so everyone can “get on the same page.”

AI‑native: Ask for a delta summary: what changed since the last reply, the decisions made, blockers, owners, dates, and the open questions that still need an answer with links back to the exact original messages as citations. If the thread implies work, extract tasks and generate an action plan.

Team of Five:

• You: What you need to do after reading (decide, reply, schedule, etc.).

• Assistant: Compile the thread(s) with links to original messages.

• Expert: Ask to flag risks or dependencies that may have overlooked.

• Coach: Track email drifts and suggest nudges to prevent them.

• Creative Partner: Draft a storyline and two alternative reply options.

Try: “Summarize what changed in this thread since Monday. List decisions, blockers, owners, dates, and the top 2 to 5 unanswered questions. Propose a 120‑word reply in my voice with 2 alternative styles, one direct and one more diplomatic.”

Success Metrics: Stakeholder rating; number of clicks avoided; % of actions with owners and dates, thread closure.

2. Open folders to locate a file

Fred‑like: Search files or attachments across folders and open them one by one to check if they’re what you’re looking for. Bonus points if you stumble upon an old file that’s not related to your search and spend a half-hour reading it out of nostalgia for the good old days. Which leads us straight to the next behavior!

AI‑native: Ask a cross‑platform retrieval with data connectors that return direct passages from all documents found on the topic and a consolidated summary of the research task, including comparing or consolidating files in case there are multiple versions.

Team of Five:

• You: Name the artifact you need (decision, number, quote).

• Assistant: Retrieve passages + links, highlight contradictions.

• Expert: Define provenance requirements (doc type, author, date).

• Coach: Note where you stored things poorly; propose a naming rule.

• Creative Partner: Propose how to synthesize the findings into a memo.

Try: “From all my folders and files since March, what is the current status on X? Quote and link, surface potential conflicts or inaccuracies, extract the latest and highlight deprecated versions of any old document version.”

Success Metrics: # of clicks avoided and minutes (or hours) saved, contradictions found and old versions archived.

3. Search the web with keywords

Fred‑like: Type a keyword that you know yields great results (after all, you’ve been successfully using it for decades now!). Bonus points if you open (and keep open) multiple  tabs to dance from one to the other like a true browser hero, skim through the content and guess as you scroll dozens of links on Google, products on Amazon or videos on YouTube.

AI-native: Share context and goals with your favorite LLM platform. That’s not a search bar, that’s a dialog box so let’s have a conversation. Request evidence tables with claims, links and dates, as well as counter-arguments from dissenters or disappointed viewers / buyers, and a short verdict on the criteria that specifically matter to you with a confidence score attached. If it’s a recurring topic, assign a task to generate weekly update briefs with citations, deltas and emerging trends.

Team of Five

• You: Define the question, constraints, and “decisions at stake.”

• Assistant: Pull top sources, extract quotes, build the evidence table.

• Expert: Specify trusted source types and define evaluation criteria.

• Coach: Track bias (recency/authority), log what would change your mind.

• Creative Partner: Suggest follow-up questions and alternative requests.

Try: “Answer in 5 bullets with links and dates; include the top counterargument and what data would change the conclusion. From this report, produce an action plan with milestones and watchouts.”

Success Metrics: Time‑to‑information; # of claims with citations; decision clarity and lack of viewer / buyer remorse.

4. Run meetings without transcripts

Fred-like: No recording, disorganized notes, a hazy recollection of events and no follow-up. You never review your own performance, relying solely on memory. Bonus points if you interrupt other participants before they finish their sentence, even though in this case you’ll probably never realize you earned them.

AI-native: Consent-based, recording is for decisions and coaching, not surveillance. The agenda is automatically generated from goals and past transcripts. Pre-reads are kept minimal, with key highlights, and presented as Q&A or podcasts. Recordings and transcripts lead to clear decisions, assigned owners, and deadlines. Attendees receive a concise brief. The next agenda for the next meeting is automatically generated based on goals and AI notes. A coaching view helps you track clarity, pacing, interruptions, filler words, and stakeholder questions, allowing you to set a single behavioral goal for your next interaction.

Team of Five:

• You: State purpose, outcomes, and what “done” means.

• Assistant: Capture transcript, tag decisions/actions, draft recap.

• Expert: Pre‑load a checklist of must‑ask questions.

• Coach: Analyze talk time, interruptions, clarity, behaviors to tune next time.

• Creative Partner: Propose next‑meeting agenda from unresolved items.

Try: “From this transcript, list decisions (with owners/dates), unresolved questions, and a 120‑word recap. Add a 3‑bullet personal coaching note and generate the next meeting's agenda.”

Success Metrics: time to recap; action completion rate; one behavior improved per week.

5. Type a first draft or an initial prompt manually

Fred‑like: Type, backspace, self‑edit, repeat. Bonus points if you type with 2 to 6 fingers because you’ve always been too busy to learn typing. The good news in that case is that you might end-up leapfrogging typing altogether, at long last.

AI‑native: Gone are the days of Dragon Dictation! Current AI technology offers exceptional accuracy in transcribing voice notes and excels at speaker identification and sentiment analysis. This enables the creation of structured initial drafts that capture much more context than you would even have typed, and can repurpose them for any target audience, including yourself if you’re journaling. By specifying the desired tone, reading level, and objective, you can even generate 2-3 distinct versions, each accompanied by an explanation and expert review. Your role is to finalize the remaining 15% of the prompt or content.

Team of Five:

• You: State audience, objective, desired length, non‑negotiables.

• Assistant: Transcribe, structure, and format; insert citations.

• Expert: Run accuracy checks and verifies linkable claims.

• Coach: Flag language quirks; suggest style cards by audience.

• Creative Partner: Offer two alternative openings and a metaphor.

Try: “Turn this 2‑minute voice note into an executive email (≤200 words) and a 5‑bullet summary alternative." or "Use it to run a Deep Research and explore the topic with citations and references. Explore pros and cons of each idea that I shared with reasoning and then use critical thinking to balance perspectives and invite me to a higher level of understanding.”

Success Metrics: draft‑to‑ship time; edit passes, surfacing and solving "unknow unknowns".

6. Write checklists from memory

Fred‑like: Freewrite a to‑do list, then hope you have not forgotten the one step that explodes the whole process. Bonus points if you recreate the same checklist every month because the last one is lost in a notebook somewhere.

AI‑native: Generate the checklist from the best current knowledge asset. It can be a Deep Research on the Internet, transcripts, emails, prior documents, or even screenshots. Then add a 60‑second pre‑flight check and the three most common failure risks right under it. For recurring work, save the checklist with roles, triggers, and “what good looks like” in an AI Notebook or Project.

Team of Five:

• You: State the outcome and constraints (time, tools, risk tolerance).

• Assistant: Extract steps from transcripts / docs; order them, add owners, etc.

• Expert: Add industry‑specific steps and compliance checks; define “done.”

• Coach: Review what went wrong before and pin those items at the top.

• Creative Partner: Propose missing steps or a lighter version for emergencies.

Try: “From these notes and last month’s recap, generate a 12‑step checklist for onboarding a vendor. Include a 60‑second pre‑flight, three likely failure modes, and bold the irreversible steps. Output as a one‑pager.”

Success Metrics: Incidents per run; time to complete; onboarding time for a newcomer using only the checklist.

7. Ideate Alone

Fred‑like: Stare at a blank page until inspiration arrives, then capture the first decent idea and fall in love with it. The workshop in your head is a very exclusive venue with one attendee and no dissent. Bonus points if someone asks you about it, you tell them it’s “not ready for prime-time”.

AI‑native: Seed your favorite AI with constraints and overarching goals, then generate diverse ecosystems of options. Ask for non‑obvious angles (perspective change, analogy, topic inversion, “what if we removed the constraint?”). Draft quick feasibility notes, and a convergence step that narrows to two paths with a simple experiment that can be done in hours, not days to A/B test them.

Team of Five:

• You: Define the core challenge, objectives and criteria for idea selection.

• Assistant: Compile initial references and comparable cases; attach sources.

• Expert: Score on originality, feasibility, etc. Suggest smart outcome metrics.

• Coach: Protect divergence time; capture bias patterns, help differ judgement.

• Creative Partner: Push beyond the obvious with “yes, and…”, reframes, etc.

Try: “Give me 10 non‑obvious ideas to increase retention for a zero‑budget campaign. For each, add a two‑line rationale, a comparable example in a different field to learn from, and a first experiment I can run this week.”

Success Metrics: Number of ideas that survive expert screening; time to first experiment; learning captured per idea. Eventually, rating across criteria and meeting the objective.

8. Improvise a tough conversation

Fred‑like: Wing it. Rely on charm, caffeine, and good intentions. Hope you remember the crucial sentence and your ideal outcome when the stakes are high and the clock is ticking.

AI‑native: Script the intent and desired outcome, list likely objections with respectful responses, and rehearse with a role‑play agent. Ask for three variants of the opener (direct, diplomatic, data‑led) and an “if things go sideways” branch. Afterward, debrief from the transcript: what you said, what landed, what to change next time. (For more, see “Train An AI Twin…”)

Team of Five:

• You: Write the single sentence that defines success for this conversation.

• Assistant: Draft the opener, objection‑responses, and a two‑minute close.

• Expert: Recommend framing and order; add risky language to avoid.

• Coach: Score clarity, pacing, and objection handling; set the goal to land.

• Creative Partner: Offer different approaches (story, metaphor, values).

Try: “Help me script a five‑minute opener for a scope‑creep conversation with my manager. Provide three tone variants, the top five objections with respectful responses, and a 60‑second close that preserves the relationship while establishing solid foundations for the future.”

Success Metrics: Outcome vs. intent; counterpart satisfaction (ask!); improvement on one behavior metric (e.g., interruptions, hedging).

Using This Habit Tracker in Practice (One Week Pilot)

Let's be honest, bad habits are hard to track and even harder to change. It's actually a miracle to do either and borderline impossible to do both. That's why we have them in the first place, even though they often used to be good habits at some point. But when we do, these little bundles of neural architecture and daily hacks can change your lives for the best! So let's genuinely give it a chance. It's all about first steps, and here's a concrete way to get started.

• Pick 3 daily gestures (e.g., mail search, meeting capture, conversation prep).

• For each, run the Team of Five explicitly once. Keep the outputs.

• Track three numbers: minutes saved, # of decisions made with evidence, and one behavioral metric from the Coach (e.g., speaking time, filler words, on‑time follow‑ups).

• At week’s end, write a 1‑page "de‑Fred" report: before/after time, quality lift, how you felt about it and one habit you will keep going forward.

Culture difference between AI‑native and AI‑reckless

Now every old habit we have is NOT a bad habit. Quite the opposite, most of them are more valuable than ever, and can help us not drift the wrong way and overindex on AI either. So here's a quick reality check to keep the wheat while we get rid of the shaft and double down on lifelong good habits which are increasingly relevant in 2025.

• Default posture: Ask, then act. Ask the AI first; require sources, deltas, and confidence; then apply judgment, close the loop with our initial intent, run ideas with partners.

• Skills to elevate, not erase: problem framing, intent and purpose (what “good” looks like), and sense‑making across disciplines. Critical thinking, constant iteration, boundless curiosity, enjoying the journey as much as the destination

• Human in the loop where it matters: manually reviewing decisions, consent, sensitive data, and public‑facing claims matters. These are not bad habits or Fred-like behavior!

• Data over intuition: track time‑to‑answer, citation coverage, draft‑to‑send ratio. be mindful both of hours saved by automation and ideas generated by ideation.

• Skills to elevate, not erase: problem framing, taste (what “good” looks like), and sense‑making across disciplines.

Conclusion: What Awaits Us on the Other Side of Fred

We stand at a unique inflection point. The choice before us isn't about replacing human ingenuity with artificial intelligence; it's about augmenting our innate capabilities by freeing ourselves from the "Fred-like" habits that anchor us in the past.

After all, my son didn't plan to spend his day on YouTube but discover Hades II Early Access!

Embracing an AI-native posture isn't about surrendering our judgment, but about elevating it, allowing us to focus on the truly human aspects of our work: framing problems, seeking deeper understanding, and forging meaningful connections.

By consciously integrating AI into our daily gestures, we cultivate not just efficiency, but also a renewed sense of curiosity and a deeper appreciation for the journey of creation and discovery.

Let's step forward with optimism, knowing that by shedding the friction of yesterday, we unlock a future where our potential is amplified, and our human spirit can flourish at last. Charlie Chaplin didn't shoot the Modern Times to praise inhuman repetitive tasks at the workplace, and I believe he would smile at reading how we're still on the path.

Related: Redefine AI: Amplified Intelligence, Augmented Intelligence
Related: Train an AI Twin to Practice Difficult Human Conversations
Related: Beyond the Prompt: How to Make an AI-Driven Culture Transformation

Join over 24,147 creators & leaders who read Methods of the Masters each week