Reclaim 6 Hours of your Week in 10 mins

Happy Monday. It’s about 9 am, your coffee’s hot, and your brain hasn’t been hijacked by Slack yet. This is the most valuable 30 minutes of your week, so let’s not waste it answering emails. Let’s set up the machine that’ll answer them for you.

The problem: Most people open a blank Claude chat every single morning and re-explain who they are. That’s like onboarding a new employee daily and firing them at 5pm. Exhausting, and you never get the good stuff.

Spend 15 minutes once this morning, and every check-in for the rest of the week takes 30 seconds. Here’s the ritual.

By default, Claude opens with "Great question! Let me break this down for you…" — three lines you didn't ask for, every time and it costs you credits.

Newer Claude models follow instructions very literally, which is a gift: tell it exactly how to behave and it sticks. Go to Settings → Profile preferences (free on every plan) and paste something like this:

Prompt:

“ Skip the preamble. No “great question,” no restating what I asked, no closing summary unless I ask. Lead with the answer directly.
When I ask for options, give me your actual recommendation, not a neutral list.
Be direct. If I’m about to do something dumb, say so.
Before you start, ask me the 5 most important questions that would help you do this well.
After I answer, then begin. "

Two rules from Anthropic’s own playbook make this stick: say what to do, not what to avoid (”lead with the answer” beats “don’t ramble”), and explain the why when it matters (”be concise, I read these on my phone”). The model generalises from reasons.

Think of a Project as a workspace with a memory , your instructions and files ride along into every chat inside it. Build 2-3 you'll open daily, and feed each one your style guide, past issues, and brand voice. Here's mine.

Project: “Newsletter.” Train it on your voice with 3–5 real samples (Anthropic says 3–5 examples is the sweet spot for steering tone).

Prompt:

" Here are 4 things I’ve written: [paste]. Don’t describe my “vibe” — extract the patterns: sentence length, how I open and close, words I reach for, what I never do. From now on, draft in this style and don’t drift back to your defaults. And don’t summarise the write up unless I ask you to. "

Now, all week: “draft Thursday’s intro” → it sounds like you, not a press release.

After being in the ecosystem for so long I have realised that underspecified chatter actually lowers answer quality. So flip it: Ask it not just the questions that confuse you but also the ones that should or already question you .

Prompt:

"Here’s my plan: Before you give your opinion on any of my ideas use this flow
Step 1: Attack it. Find every assumption that could be wrong and every realistic way it fails. Don't be polite.
Step 2: Steel-man it. Make the strongest possible case that I'm right.
Step 3: Tell me what you actually think, and what you'd change after considering both the ways it could go very wrong and slightly right."

Two minutes answering > twenty minutes correcting. Every time.

When you paste a long doc or dataset, put it at the top and your question at the bottom. Anthropic’s testing found query-last ordering can improve answer quality by up to 30% on big inputs.

These are a few content pieces that I personally enjoyed and I think would be a great watch for those 3pm slumps , lunch breaks, slow evenings, or the small pockets of time between everything else. And at the end, there’s also a playlist to get your week started with a little more energy and momentum.

So here are a few things worth carrying into your week:

1. “How to Find Work You Love”— Scott Dinsmore’s TED Talk, an 18 minutes talk that might genuinely rearrange your week. He spent four years studying why nearly 80% of people don’t love their work, and what the other 20% do differently. Watch it, then ask yourself whether your Monday to-do list reflects the answer.

2. Two Minute Papers — bite-sized, genuinely delightful breakdowns of the latest AI research, delivered with the most infectious “hold on to your papers!” energy on the internet.

3. Simon Willison's blog — if you read one AI engineering blog, make it this one. Daily, hands-on, zero hype. He builds the thing, breaks it, and writes up exactly what happened — prompt-injection war stories, new-model tear downs, tools-that-actually-work. It's the closest thing the field has to a builder's lab notebook.

4. Andrej Karpathy's nanochat — Karpathy released a full-stack training and inference pipeline for a simple ChatGPT-style model — roughly 8,000 lines of clean PyTorch spanning the whole pipeline from data prep to deployment. He calls it "the best ChatGPT that $100 can buy."

5. How to Work and Compound with AI — His core move: treat verification as a ladder — catch errors at the cheapest, most deterministic rung before they reach you. Context as infrastructure, taste as config, scale via delegation.

6. Spotify Playlist — I’ve actually been enjoying this playlist lately and it’s perfect for morning runs, work sessions, gym days, or just getting through errands with a little more momentum.

Three teams building at the edge, looking for builders this week:

• Smallest.ai — building ultra-low latency voice AI systems for enterprise-scale speech and multilingual intelligence.

  Open roles span Forward Deployed Engineering, Growth, and AI infrastructure→ https://jobs.ashbyhq.com/smallest.

• Genspark — an AI-native search and agent startup founded by former Baidu leaders, building multi-agent AI systems to rethink search and workflows. Roles are listed through Gem, with teams growing across Palo Alto, Singapore, and Seattle, explore openings via their careers portal - https://jobs.gem.com/genspark .

• Eragon AI — building company-specific AI systems that continuously improve from real-world usage and feedback loops. Hiring aggressively for Applied Research Engineers, ML Engineers, and Technical Staff in San Francisco, apply through their careers page - https://www.eragon.ai/careers

I hope this week is productive in the ways that matter, not just efficient, but meaningful. The goal isn’t to exhaust ourselves chasing potential, but to grow into it without losing our peace along the way.

Have a lovely week ahead.

— Team AI+