CEOs that operate with excellent analytical observability

Traditional companies trying to compete with AI-native start-ups.

Finally figured out what LLMs are aspiring to sound like. It’s @AdamMGrant

A trust-reducing maneuver of people who are uncomfortable being wrong: they struggle to lead with an opinionated headline. They’ll dive straight into descriptive facts instead: “This candidate spent 10 minutes on the first problem and needed multiple hints…” “We had 5% week on week growth…” “We explored 7 design variations of this problem…” No no no! What the other person ACTUALLY wants is for you to earn their trust and respect their time. ”I’m a solid no hire.” “We had okay but unspectacular growth last week.” “I have a design solution I’m happy with.” THEN, follow-up with your evidence / context / facts. Sure, you may be wrong. The other person might hear your evidence and say “I actually think we had a spectacular week” or “I don’t think this proposal cuts it” but actually having these conflicts to calibrate mental models is PRECISELY what builds trust over time. If you’re uncertain, you can contextualize your opinion (”I’m still new to recruiting but I’m a solid no hire on this candidate…” or “I might not be calibrated yet but I have a design solution I’m happy with…”). But please do not just state the facts. That’s your insecurity talking. Make like a journalist and lead with the headline.

Signs of engineering competence (that even non-engineers can sniff out): 1. When asked to explain a complex topic you know little about, does their explanation make you feel confused / dumb or way smarter afterward? 2. When confronting a bug, is it case closed when the bug is gone, or do they dig like a dog on a bone until they find the systemic root cause(s)? 3. When faced with major technical decisions, do they rely on social proof (Youtube / forums / friends), or are they also reading the full documentation / relevant papers? Independent, clear thinking is getting harder and harder to find, especially in the era of AI!

The market for start-up design talent has never been more competitive. I’ve had half a dozen companies ask me recently, “how do I get a great founding designer?” Folks, if you are not already bosom buddies with great designers, you’re gonna need to put out a beacon that attracts top talent: 1. Tell a compelling, true story about your ‘why.’ Make a page dedicated to this. Make your story come alive as only you can. Anyone with ChatGPT can write a generic sounding req about competitive benefits and rocketship grow opps. Write about a problem and vision that is wildly ambitious yet deeply personal to you. 2. Express that you “get” the importance of design. This is your ‘proof of work’ that you will value the person that you hire, for ex (not limited to): a. Your page / app / website show attention to detail and care for the experience b. Your description of the job shows a deep understanding of what good design means (do your research here). c. You talk about design work that you admire or find aspirational. d. You have design-forward people on your cap table (ie @designerfund, @soleio, @karrisaarinen, @rsg, @tobi, etc) Ofc easier said than done, but the key is to put some effort into standing out as an incredible environment for designers to do their best work.

What iteration to PMF looks like.

In college I met this guy at a social and we split off from the group to spend 3 hours talking about fave video games / books like Final Fantasy and Ender’s Game. Years later we got married, had 3 kids and gave them names from both the Final Fantasy and Ender’s Game universe.

Why do all saas websites grow less creative and more boring over time? It used to puzzle me until I went through it myself: 1. 0-1: your website tries to appeal to top talent. Creativity stands out. 2. 1-10: your company tries to appeal to early adopters. Creativity stands out. 3. 10-100: your company tries to win over people who wear collared shirts / dress shoes to work. You match what they know and trust.

What stops LLMs from getting smarter by giving higher training weight to the inputs and feedback from people who think more critically? @grok has a nice advantage here in its interest graph.