Alex Thielman

Fractional CTO • Cofounder of Tech Leaders • Dad👨‍👧Drummer 🥁MTBer 🚲

Most AI advice is noise. After helping hundreds of individuals and organizations navigate AI adoption as a fractional CTO and founding engineer of Digit (acquired by Oportun), I've seen what actually works (and what's just hype).

Your biggest risk in 2026 isn't AI. It's getting distracted by shiny objects while your competitors ship more value to customers, faster. Evolution wired us to fear missing out, but that same instinct makes us chase every new tool instead of building what matters.

We help technical leaders and their teams build custom AI workflows that deliver real ROI, without the fear, uncertainty, and doubt. No complex SaaS. No messy "vibe coding." Just proven playbooks for AI transformation.

Whenever you're ready, here's 3 ways we can help you with AI:

Watch some of our best trainings for free on YouTube here: https://www.youtube.com/@aitechleaders
​
We've helped 200+ indie consultants, SMBs, and leaders from organizations like Amazon, EPA.gov, Vercel, and LegitScript grow with AI. Book an AI Strategy Session to see if we can help you with a customized AI workshop, training program, or speaking event: https://www.technical-leaders.com/ai-call​
​
We’re about to open up limited spots in our AI SOS™ program. Go here to be notified first: https://www.technical-leaders.com/ai-waitlist

P.S. Send me a direct message, I love hearing about what you’re up to with AI

Todd Larsen

If AI coding tools have disappointed you, you're probably missing one thing.

→ Context engineering.

Most developers install Copilot or Cursor, type a prompt, get mediocre code, and assume the tech isn't ready yet.

But the tool isn't the problem.

The setup is.

Here's what actually works:

→ Use Claude Code running Opus 4.5 directly. Not through an IDE layer. The difference in output quality is significant.

→ Feed it structured context. Project architecture, coding standards, relevant files. The more it understands your codebase, the better it performs.

→ Treat it like a senior developer who just joined the team. It's capable, but it needs onboarding.

I've been running this setup for months.

The velocity gains are real. 10x isn't an exaggeration.

Most technical people gave up after one bad experience with AI coding.

The ones who figured out context engineering?

They're building faster than they ever have.

Ex-Digit Founding Engineer → Helping Non-Technical Leaders Build AI Workflows Without the Hype • CTO @ AcreHedge

How to get 10x better AI answers

(works with any AI model)

Turn on “Atomic Thinking” mode

Just copy+paste this text at the beginning of your AI system prompts to shift from the default “Linear Thought” mode (which is super limited and narrow):

—

Instead of Chain of Thought: "Solve this step by step" approach, use the Atom of Thought approach like this: 

Decompose this problem into atomic reasoning units. 
For each atom:
1. State the logical component 
2. Validate independence
3. Verify correctness 
Then synthesize atoms into final answer

Custom GPTs can be limiting. Skills offer a more advanced – and accessible – approach to AI agent workflows.

This video explores how skills, developed in part by Anthropic, can be leveraged for enterprise adoption, creating more robust and adaptable AI solutions.

What if 'skills' are the key to unlocking AI's full potential in the workplace?

taplio

The Battle

Alex Thielman

Todd Larsen

Followers

Posts

Engagements

Likes

Performance ScoreTodd Larsen's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Alex Thielman

Todd Larsen

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Todd Larsen's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreTodd Larsen's LinkedIn Performance