Rob Go

Lucky husband of @ntduke and father of two amazing girls. Cofounder of NextView. Built some product at Ebay and learned some investing at Spark Capital

[Rob Go @robgo]

@VCnerd_ I’m similarly excellent at finding things in the heavy rough

[Rob Go @robgo]

There are basically three approaches to venture. And one is dominating the attention, headlines, and capital.

1. Fighting for diamonds in the middle of the fairway
2. Searching for diamonds in the rough
3. Mining for silver

[Rob Go @robgo]

One misconception of this is that lower price does not = diamond in the rough. The market is actually quite efficient. A diamond in the rough can still be priced at $60M pre, if it would have otherwise been priced at $120M pre with a more consensus fact pattern.

[Rob Go @robgo]

There are basically three approaches to venture. And one is dominating the attention, headlines, and capital.

1. Fighting for diamonds in the middle of the fairway
2. Searching for diamonds in the rough
3. Mining for silver

[Rob Go @robgo]

@ewiesen I know at least one of my colleagues would say exits are the biggest problem, but yet.

[Rob Go @robgo]

@jamesflynn_2 That's true almost by definition, unless you're one of a very small number of next-gen firms with top 00001% access.

[Rob Go @robgo]

@bznotes Yes, it has always been this. The relative attractiveness of each has definitely shifted dramatically over the years.

We generally don't mine for silver, but I actually would love to see that ecosystem take shape - I think it's a real gap for founders.

[Rob Go @robgo]

@chudson @miafarnham Really enjoyed it. Thanks for having me! It actually spurred me to think about a couple follow-up posts that I'll try to write over the next week or two. Thanks again for such a great podcast!

[Rob Go @robgo]

@jeffwhelpley I really want this to work. But there are a lot of challenges to this approach. It works in growth equity, but questionable before PMF.

[Rob Go @robgo]

The most promising returns are in #2. But this is harder than ever given extreme downstream funding risk, elevated entry prices, and limited exits outside of consensus winners. Funds pursuing this need to consider revamping their approach entirely. The old model won't work.

[Rob Go @robgo]

#1 has been dominating for many reasons. Lots written here.

#3 is funding companies on a non-traditional venture path. A worthy effort, but the question (as in silver) is whether you can really do this profitably.

[Rob Go @robgo]

Great overview of how to think about security risk for Muse. Hilarious last post on instinct at the bottom of the thread. YOLO!

[Melody Koh @melodykoh]

As part of testing @Muse, I quizzed it to explain to me a few things I think about when using these agents and deciding how much connections and context to give - the struggle of the Trust-Utility Curve (the more you give, the more useful it can be). https://t.co/tQsBErXpTE

TLDR - there continues to be limitations around how LLMs fundamentally work and the probabilistic nature of things, and there really isn't a way to get around on context leakage or prompt injection-like attacks. However, I found Muse/Meta's willingness to be transparent in its mechanics reassuring.

A few highlights on my Q&A with Muse - see more details on the screenshots of my questions and its responses. 🧵

Q1 - you say that you will ask for my permission for actions, but how do I know this can be enforced deterministically (vs instructions in files, which are just vibes for these models)?

Most reassuring part of the answer:

"When I invoke a send tool, the runtime intercepts the call and blocks execution until you've approved it. I can't trigger the approval myself, can't see it, can't override it — the approval card is invisible to me and your decision is final. So even a compromised or injected version of me, actively trying to send without asking, hits the same wall: there is no code path where the send executes without the runtime obtaining your decision first."

Translate to non-developer speak: there's a mechanical gate that forces the model to serve up approval to the user when the action is a list of predefined "must get user approval". No vibes-bsed (good!)

[Rob Go @robgo]

Had an awesome time hanging out with @yifanz at the AI House in Seattle. They are building something really special there. Really excited for demo day - good luck!

[Rob Go @robgo]

@credistick VC's with false confidence? Say it ain't so!

[Rob Go @robgo]

@credistick I do think that when used well, it leads to much faster decision and much sharper conversations. Using it well means leveraging claude to augment human thinking vs. turning over the reasoning and judgement altogether.

[Rob Go @robgo]

@howardlindzon Happy Birthday!

[Rob Go @robgo]

@ChristinaPhili5 great mantra!

[Rob Go @robgo]

Congrats! Pretty much a perfect and obvious fit all around.

[Rebecca Kaden @rebeccakaden]

https://t.co/axWkYS2z4E

[Rob Go @robgo]

@gdibner @nikunj @whoisnnamdi @ethankurz very clearly written - great post!

[Rob Go @robgo]

@credistick The challenge for your 2nd point is that the relatively smaller exits have not been very plentiful. And the entry price for all early stage has risen so dramatically that these exits still need to be very big to drive returns that are superior to the megafunds.

[Rob Go @robgo]

Or maybe it's just a healthy check and balance. In a perfect world, compounding illiquid investments is not a problem. But there is so much asymmetric info and misaligned incentives that pushing for DPI alleviates some of this.

[Rob Go @robgo]

I had a thought. Curious what LPs think. cc @endowment_eddie

Do LP's actually want DPI? If a manager is one you love, you like the remaining investments and you believe their performance will persist, aren't you more than happy to have value compound?

[Rob Go @robgo]

You'd rather have DPI only if a) you believe that the forward looking IRR for the fund is below your targets, b) you worry that TVPI is above actual value, and you want certainty or c) you need liquidity (for actual cash needs or to match some asset allocation model)