Session brief
How to Fail at AI Strategy: Hamel Husain & Greg Ceccarelli
Overview
This talk, presented by Hamel Husain and Greg Ceccarelli, offers a satirical guide on how to spectacularly fail at AI strategy. Instead of best practices, the speakers advocate for "worse practices" to ensure projects are torpedoed and teams are alienated. The core thesis is that by inverting conventional wisdom and embracing division, ambiguity, and a focus on tools over processes, organizations can achieve complete AI failure.
Who should watch
- AI Engineers
- Product Managers
- Builders looking to understand common AI strategy pitfalls
- Leaders aiming to avoid organizational dysfunction
Key takeaways
- **Divide and Conquer:** Create silos between teams and incentivize secrecy, preventing knowledge sharing.
- **Embrace Wishful Thinking Promises:** Overpromise AI capabilities to customers without focusing on feasibility or details.
- **Ignore Cost-Benefit Analysis:** Purchase the most expensive infrastructure without evaluating ROI, leading to technical debt.
- **Define Strategy with Ambiguity:** Use vague, all-encompassing goals like becoming the "global AI leader in everything" without clear definitions.
- **Communicate with Jargon:** Drown teams in technical jargon to obfuscate the actual work and exclude non-experts.
- **Focus on Tools, Not Processes:** Throw new tools at problems without analyzing root causes or customizing solutions to business needs.
- **Blindly Trust Metrics:** Adopt generic, off-the-shelf evaluation metrics without customization or cross-checking with domain experts or users.
- **Avoid Data:** Discourage looking at data, both personally and within the organization, by using complex systems inaccessible to non-engineers.
Notable quotes
*The point is to communicate in such a way that nobody understands.*
*Embrace Perpetual beta; just create a massive backlog in GitHub.*
*Adopt a mindset that evals are a vendor problem.*
Unofficial community note. Prefer the recording for nuance.