francis zhan

vc at @tribecap by day, game dev by night

views here are my own

[francis zhan @franciszhan]

https://t.co/LTzn5uT155

[francis zhan @franciszhan]

most consumer ai are fighting on capability, on being the all-knowing assistant, but where are my fun ais? why can't i raise a pokemon and evolve it yet? send it to parties with other pokemon? anyone building this?

[francis zhan @franciszhan]

@jonah_b love the chess analogy. even sports to an extent shows this. humans invented many ways to travel fast, but we still appreciate olympic 100m sprinters and people dedicate their lives to it

[francis zhan @franciszhan]

proud to be a continued supporter of the vision!

[Predicate @0xPredicate]

We are thrilled to announce our $8M Series A to connect regulated markets to onchain finance.

Predicate enables digital assets to reach every market compliantly. We do this by embedding compliance controls directly into digital assets to block high-risk and sanctioned activity at every transaction.

To date, Predicate has:

- Screened $1B+ in transaction volume
- Evaluated 6M+ transactions against custom compliance policies
- Blocked 100K+ non-compliant transactions

The round was led by @bcap with participation from @Paxos, @1kxnetwork and @tribecap.

[Predicate @0xPredicate]

Filing a report after an exploit doesn't stop illicit crime.

In @AmerBanker, @pumpernikhil argues banks should get credit for blocking suspicious transactions before money moves, not for just filing reports after the fact.

Read the op-ed here: https://t.co/92p28jgVwF

[francis zhan @franciszhan]

the AI boom reminds me of the moment you unlock robots in factorio on your first playthrough. intricate conveyor belt wizardry immediately goes out the window and your default solution for anything becomes plopping down roboports and logistic chests

[francis zhan @franciszhan]

business model #2/3 spotted in the wild! @trybountyai

[francis zhan @franciszhan]

been reading about the dot-com bubble and wondering what we’ll look back on as obviously wrong about this AI cycle.

my bet: the business models are wrong

five ways this could play out ↓ https://t.co/LnIk6g08VM

[francis zhan @franciszhan]

@robbiepetersen_ agreed! though I think there’s scenarios where a “curator” app or aggregator exists too, especially for domains where users aren’t sure what they want e.g. fashion. Agent has some vague intent and preferences (street wear) and the curator fulfills it against brands

[francis zhan @franciszhan]

@signulll totally agreed, people overfocus on raw capability and underfocus on the consumer ui/ux and how picky a consumer can be on average. they want out of the box only.

https://t.co/yMSM5CATXb

[francis zhan @franciszhan]

https://t.co/JdiRK9OM0r

[francis zhan @franciszhan]

@azhangtaco too scared to listen to the audiobook lest "goddammit donut" is permanently stuck in my head

[francis zhan @franciszhan]

I've always held the belief that the highest signal thing you can get from a person is all the books they’ve read. In fact, it is a stronger and richer signal than their political leaning, their religion, upbringing, etc. (even if it's fiction!!)

I'm inspired by Ben to share something similar, with a twist. Instead of favorites or life changers, I wanted to share a few books/essays I think about at least a few times every year since I read them.

Fiction:
Tomorrow, and Tomorrow, and Tomorrow (Gabrielle Zevin)
World War Z (Max Brooks)
There Is No Antimemetics Division (qntm)
Children of Time (Adrian Tchaikovsky)
They’re Made out of Meat (Terry Bisson)

Non-fiction:
Never Split the Difference (Chris Voss)
Superforecasting (Dan Gardner and Philip E. Tetlock)
Predictably Irrational (Dan Ariely)
Kitchen Confidential (Anthony Bourdain)
The Innovators (Walter Isaacson)
The Omnivore’s Dilemma (Michael Pollan)

[Ben Roy @benroy]

I think this is the most personal thing I've ever shared on the internet

It's wild to go back through the list and think man so much of who I am and how I think about the world keys off these books and essays

[francis zhan @franciszhan]

@andreistoica_ three body was THIS close to making the list for me too

[francis zhan @franciszhan]

@andreistoica_ im always eyeing your goodreads updates 👀

[francis zhan @franciszhan]

@benroy inspired by the best!

[francis zhan @franciszhan]

glad you enjoyed!
my two cents, i think key is setting expectations clear in the first place that you ARE specialized, be it product branding or onboarding. Siri is a fun case study to look into.

their original messaging pre-apple promised getting tickets, finding restaurants, and calling taxis, not much more. worked wonders, everyone loved it. Apple pushed for them to become the do-everything assistant post acquisition and you know how that went

[francis zhan @franciszhan]

https://t.co/JdiRK9OM0r

[francis zhan @franciszhan]

been thinking a lot about trust in agents. I'm a pretty experimental guy when it comes to using new products, but even I find it hard to trust an agent with my stocks account, regardless of the amount of "guardrails" in place. it's akin to how you're still scared of heights when looking over railings on mountains; logically it hasn't failed for millions of other tourists, but WHAT IF... (this extends to machine trust in general re: the autonomous car debate)

If no amount of technical guarantees work, the natural conclusion is that machine trust can only be built socially. in many ways this aligns with our psyche dating as far back as hunter gatherer era, when you were able to trust others to forage random mushrooms for you with very few proof points regardless of if you could yourself identify them.

Instinct did a great job with this by using a referral only network, and most people didn't hesitate to share credit cards with the agent as a result. But I can also see other forms of social trust working: something as simple as an existing IP lending its authority (JARVIS manages your wallet!)

[francis zhan @franciszhan]

@cremieuxrecueil would love to learn more!

[francis zhan @franciszhan]

nathan for you felt unrealistic until i remembered people take business advice from chatgpt