Miles 𓂀

finally doxing my polish ass as the founder at @HorusGroup_ across 3 operations dedicated to tech startups, we run: GTM + fundraising with $55M raised to date, AI ops for service providers, and grants. leave a follow as the team will be pushing previously gatekept gtm alpha

@brettcalhounn i like to say the founders with the ugliest decks and the happiest customers raise faster than the ones with perfect slides and no retention. but yeah working on the deck feels safer than chatting with potential and existing users

Your unrelenting desire for the perfect pitch deck will be your downfall. Be more concerned with not how polished your story sounds to investors, but how well you treat your customers.

@kevinrose works until you need distribution or enterprise sales or regulatory navigation. the AI-native model is great for building. it doesn't solve the parts of the business that aren't building (at least yet)

Old model: Give 20% to VCs at Series A, spread a 10-20% option pool across 30+ heads - each person gets 0.25-1.5%. AI-native model: Skip the round entirely. Founders hold majority. If you need more folks, vest them at 3-5%+ each. Smaller teams, bigger stakes.

@0xNairolf building for degens is easier because they already know how everything works. but then ure competing over a piece of a (shrinking) pie with projects x10 your size.

everyone building for degens while the biggest audience is actually non degens mmmmmh

@auren best teams ive worked with always realized that in 6 months theyll cringe looking back at what theyre doing today

about half of the best seed investments end up being pivots. the company that was funded becomes a completely different company. which means seed diligence on the business is mostly theater. the only thing that matters is the founding team.

@credistick the exit problem breaks everything else. when companies stay private for 12 years because there's always another growth round available, the entire chain backs up. early investors can't return capital, LPs lose patience, and the next fund gets harder to raise.

Venture capital has outgrown its ability to competently manage capital. The magic of VC is the interface between GP and entrepreneur; making judgements about ideas and people that stretch into the future. The desire to scale VC into an asset class has undermined that discipline, as the structures that enable scale have obscured idiosyncrasy. This has observable, measurable consequences. Slower innovation, weaker companies, and slipping returns. A growing desperation expressed in trying to extract more from less. I sympathise with the honest techno-optimists who believe that more capital means greater acceleration. But that bullish sentiment is being exploited by rent seekers, enabling misallocation and technological stagnation. In truth, venture capital can scale, but to remain productive it must be scaled proportionally along the right dimensions, without compromising the fundamental mechanics of capital coordination and liquidity. 1) Venture capital must not become as top-heavy as it is today. First-check firms provide the discovery of new opportunities. If the downstream market grows out of proportion with that discovery layer, everything begins to crumble. 2) Exits must not be delayed in order to absorb more capital. Going public is extremely beneficial for innovative companies, and has positive externalities for innovation generally. Private capital feels easier, but it is poison in the long-run. With this in mind, I propose four pillars of scaling venture capital toward greater productivity, in the context of national capitalism — how states can wield capital for industrial growth: The first pillar is to remember that venture capital is an exit business, not an endlessly printing markups business. Companies should be oriented towards an exit once they reach an appropriate scale and are sufficiently derisked. Historically, that has been somewhere between 6–8 years. It may be longer for others that require it (see: SpaceX), which is fine. In practice, that means not wilfully shovelling growth capital into businesses that would otherwise be public. Which means finding a more productive purpose for that capital, which may be challenging for large, lazy allocators. (Building on themes explored in a large body of research, cited in Hitting Escape Velocity.) The second pillar is to ensure that the foundation of small and emerging managers is robust, producing a healthy stream of opportunity. This runs contrary to larger VC incentives and LP bias toward brand power, but the alternative is concentration that rots returns and consensus that rots innovation. (Building on work by Martin Aragoneses of INSEAD and Harvard University’s Department of Economics, and Sagar Saxena of the University of Pennsylvania.) The third is to ensure R&D intensive technologies have access to patient early capital that can carry them through to commercialisation. This helps prevent VC simply flowing down the path of least resistance to scalable software slop. Where venture capitalists do not quite have the courage to back novel “HALO” technologies from inception, they may need outside support. (Building on research by Kyle Briggs of the University of Ottawa Department of Physics.) The fourth is to provide access to well-structured mezzanine financing for companies with extreme setup costs, from nuclear plants to clinical trials. This gives early VCs the confidence to invest in these categories knowing there is downstream capital and liquidity when an IPO may be too distant and too risky. (Building on work by Andrew Lo, of MIT’s Department of Financial Engineering.) So, in the spirit of Kyle Harrison’s techno-solutionist commitment, here is how we may address those pillars…

@neuralunlock compliance and privacy are also where most current chains are weakest. the ones with a head start in TVL built for degens (super satured mindshare-wise), not for institutions. different customer, different requirements. might not transfer at all.

The battle to win institutional market share is still anybody’s game. Some chains have a head start in metrics like TVL, outstanding loans, or DEX volumes, but that matters less than you think. What institutions care about most: compliance, privacy, and lindyness. Right now, onchain activity is mostly tokenized money market funds, treasuries, and some repo activity. The lion’s share of interesting DeFi is still on the horizon.

who's in Cannes for ETHCC and looking to raise? hmu

@andrewchen the part people miss is that "fewer people per company" times "way more companies" could easily mean more total jobs. just distributed differently. the transition is the painful part, not the end state.

on the “will AI kill jobs?” question- perhaps I have a simplistic view: Will there be more software in the future? I think yes. Too many people hated previously. Thus will there be more products in the future? Yes. Will there be more companies? Yes. More founders? Yes. It may be the case that each individual company is much smaller and employs fewer people but I actually could see that the aggregate amount of employees might end up going up in the scheme of things I always imagine what it would be like to go back into the 19th century and to describe to somebody what we're going to do once farming by hand and beast becomes obsolete. Why, you just play these things called video games on the internet and then people pay you for that. Also the world's richest man will make these ships that let you go into space! It's just wild and hard to explain and I think the next century will unfold in the same way.

@KhanAbbas201 brand accounts were already fighting uphill against personal accounts in reach. this just makes it official. the workaround is the same as before: have real humans post from personal accounts and let the brand account be the landing page.

I think a lot of brand accounts will strongly get affected by the new algorithm given they are not constantly engaging and tweeting like personal accounts. The new algo tanks your engagement if you're not active 24/7 and I'm seeing very good brand accounts pushing good content get 200 views on their campaigns even if they're subscribed to the gold plan. The new algo is only for the terminally online.

@sjdedic the boredom phase is also when the best deals happen because nobody's competing for them. founders with traction right now have zero pressure on valuation from other investors.

When it starts to feel more exciting to follow the S&P 500 or oil futures than crypto, something is very wrong. Or very right. Because every time this happened before, it was the exact moment before everything changed. And it makes sense. This is when the last speculators finally tap out. The airdrop meta is dead. The memecoin casino closed. The KOL deals dried up. There’s nothing left to extract, so the extractors and shillers are gone. Just sideways charts and a timeline so quiet it makes you question why you’re still here. And I get it. I feel it too. No one is going to pretend this isn’t tough. But this is also exactly where every major cycle started. Not in a crash. Not in drama. In boredom so deep that even the strongest believers who built this industry started doubting. But that silence has never lasted. And those who gave up never came back in time. If you’re still here reading this, you’re either going to be the greatest winner or the greatest loser. But at least you’ll be great. Not mediocre like everyone who left.

@joeljohn number 3 is why most crypto projects with great research blogs and no users eventually run out of runway. the market doesn't care how well you can explain your architecture.

Some random things I’ve been thinking of 1. Markets are the ultimate judge 2. They are more permissionless than most protocols 3. Traction beats distributed design systems and essays 4. Because it compounds labor, capital to a shared cause.

@jbrukh both are bottlenecks but at different stages. right now most teams can't even get agents to do anything useful enough to generate data worth training on. the data problem comes after the utility problem.

I strongly disagree with this. The real bottleneck is generating enough agent data to be competitive.

Tim Draper passed on us. two years later, he backed us. what changed? >we didn't pivot. >we went back to the same investor with proof instead of a deck. the word "no" from the right person means "not yet." learn to hear the difference.

true though depends on the use case. some institutional stuff just won't happen without permissions. the question is whether the permissions sit on top of an open network or replace it entirely. things might need shortcuts to get anywhere - the die-hard private/decentralized solutions i got into web3 for failed to scale sadly.

If crypto wins the institutional TAM with permissioned, closed-source networks, then that is not winning at all. We didn’t come this far to take the shortcut with centralized databases. In addition to performance and scalability, the winning protocols must be censorship-resistant, secure, verifiable, and private.

@andrewchen another one: does the team's biggest expense line item go to inference costs or to the marketing site. that tells you everything about where the AI actually lives in the product.

how to tell the difference between AI-native products versus when AI is bolted on after the fact... fake AI products: - main AI feature is an AI button with sparkle icons - chat pane where you can ask LLM questions - no memory/personalization beyond one chat - users try it once and go back to using the app the "normal" way - AI is optional not essential to the product working AI native products: - you can spend $100 or $1000 via tokens as you use the product - it gets substantially better every 6 months as base models improve - core workflow is impossible without AI, not just enhanced by it - creates behavior change when users try it what else should be on this list?

@negacz klasyk, chociaż następna fala tego samego pójdzie w automatyzację. firmy będą skalować procesy które nie działają i dziwić się że AI nie pomogło.

“Wdrożyliśmy CRM”. “I jak?” “Handlowcy go nie wypełniają”. “To po co wdrażaliście?” “Żeby mieć dane”. “I macie?” “Nie”. Wdrożenie CRMa bez procesu sprzedaży przez firmę, która specjalizuje się głównie w oprogramowaniu, to kupienie Excela na sterydach za ciężkie pieniądze. W ostatnie 7 lat audytowaliśmy ponad 1200 polskich firm. Źle wdrożone CRMy to moim zdaniem najgorzej wydawane pieniądze w Polsce na technologię.

@AlmostMedia the thin skin part gets worse after the first big win. suddenly every disagreement feels like disrespect. hardest founders to work with succeeded once and never got challenged again

I love startup culture with all my heart but the one tricky bit of working with awkward outsiders who gain wealth & power is that we have a disproportionate number of people with paper thin skin

@ajhodls not many but its gonna get worse. plenty tier 2-3 names are struggling to raise and/or pivoted towards general fintech/ai

how many 2019-2021 vintage crypto funds do you think raised fund 2

6 months. 20-30 VC calls after pulling in all the favors and intros. zero term sheets, round's dead. heard this three times this quarter; good teams, good products. all cooked themselves the same way. before you open your round, ask: it's 6 months from now and we haven't closed. what killed it? it's always one of these three. you're pitching people who can't say yes 20 conversations sounds like progress. pull up the list and it's all tier 3 funds who couldn't lead if their life depended on it. they're waiting for someone else to go first. just like you. 6 months of energy into the wrong room. you don't need more calls. you need 5 conversations with GPs who actually write lead checks at your stage. the rest is noise until that's locked. nobody's afraid of missing you no partnership going live, no usage moving. no other fund circling. nothing that makes "let's revisit next quarter" feel like a mistake. so VCs keep taking your calls. keep saying "super interesting, keep us posted" and you keep confusing that with progress. watched a project go from $2M "committed" to zero in one week. lead fell through. every soft commit vanished same day. no urgency means no round. just theater. your pitch dies in the partner meeting you crush the call with a GP. doesn't matter. monday morning they walk into a room full of partners who've never heard of you and try to re-pitch your project in 60 seconds between two other deals. if they can't do it in two sentences you're out. most founders obsess over their deck. wrong thing. obsess over the sentence your champion says when you're not there. "they do X, already have Y, and Z is live next month." if a GP can't say that from memory you gave them the wrong story. --- run this exercise before you start raising. not 5 months in when the runway's gone and the team's fried. been thinking about what "fundable" actually means lately. not the word. the work. more on that next week.