Every Big Company is Focused on AI

Originally posted as a Twitter thread on September 12, 2023


1998: “The Internet is stupid, people won’t buy X over the internet”
2008: “The iPhone is stupid, my BlackBerry works fine”
2010: “Cloud is stupid, on premise is more secure”

BigCo myopia created opportunities for startups.

But today, every BigCo is focused on AI:

This is why the “known unknowns” and “unknown unknowns” make for much more interesting markets:

https://a16z.com/financial-opportunity-of-ai/

When to Escalate vs Wait

Originally posted as a Twitter thread on September 01, 2023


When to Escalate vs Wait – implications for M&A, deals, dumb policies, etc

When an outsider presents an organization with evidence of “complex wrongdoing” (or mistakes) from within — where I define that term as a form of wrongdoing that *requires* internal corroboration — almost inevitably internal antibodies form to fight off the foreign accusation.

My general theory is that organizations optimize for internal harmony — not shareholder value or customer satisfaction. The CEO will defer to his or her VP, who will defer to his or her Director, etc. “We made a mistake — we should re-evaluate!” rarely comes up because the CEO is unaware of the particulars and policies.

This happens all the time in schools with placements, companies where egregious errors get committed (eg M&A diligence), investment firms where termsheets get pulled, Covid policies that made no sense, etc. Escalation almost never works because of specialization. Very different from “clear to anyone” wrongdoing/mistakes — “that person shot me, here’s a video!”

There’s also a seemingly innate human instinct to not admit a mistake – “saving face” is an almost universal human desire. But within an organization this instinct almost metastasizes and ossifies positions from bottom to top.

I’ve written extensively about TrialPay’s M&A travails:

https://x.com/arampell/status/1562557861145636866?s=61

In one case, an M&A process died when an engineer who didn’t understand our tech said it was “bad” — and subsequently left the company to start a competitor! Everything about it was ridiculous. https://x.com/arampell/status/1562557861145636866

But my mistake (even though I was right!!) was angrily escalating to the top — thereby releasing the full antibody response and ossifying the company’s viewpoint. The wrongdoing and mistakes were simply too complex to present without internal corroboration, which settled down to the same engineer.

I also learned that more data doesn’t really do anything to change most people’s minds. Time is more important. For some reason, humans are capable of changing beliefs when enough time has passed — not when contradictory evidence emerges. You really need both.

More on this here, from my lessons in (ultimately, after much trial and error!) selling my company: https://x.com/arampell/status/1610761687547940864

Key learnings:
-time > data. It’s frustrating but patience often beats action.
-escalation normally ossifies positions — it’s much harder to fight a strong antibody response than a weak one
-if escalating, make it about the principle – or something that does NOT require internal corroboration of the specifics. How does matter to the top / higher person in a way that doesn’t require looking up the minutia?