Summary

Jordan Nel, ex-Hummingbird/Nomads, refuses to defend venture as an asset class: “I’m not sure that venture beta justifies the illiquidity premium.” His case is strictly bottom-up — “for every outlier company, there are at least three outlier funds”: every Coinbase has a Ribbit, a USV, and a dozen funds that return multiples off one name. Even in 2021 there were pockets of arbitrage (“if you were fishing in Indian life sciences, probably you’d be the only one”), and the LP’s job is finding those pockets, not timing beta.
Emerging managers shouldn’t try to beat Andreessen or Sequoia — they should see companies first and “act as signals to the bigger funds such that they will eventually follow on from you.” Big-fund GPs on 20-person teams aren’t incentivized to fish for a $5M post-money-cap founder in the backwaters; they need to deploy $30–100M tickets to drive carry. It’s not either/or: big funds run fund-of-funds programs that seed emerging managers and use them as deal flow.
Nel rejects “contrarian” for “pre-consensus” — the art is making founders “legible to capital” and sequencing the cap table. Bridget Mendler (Disney child star turned PhD, founder of Northwood Space) could have raised from megafunds but deliberately built her pre-seed with signal angels; a founder from Abuja, Nigeria couldn’t get a term sheet for six months until one risk-on pre-seed GP unlocked the San Francisco hard-tech ecosystem. On a founder with direct access to top funds, Nel says he would do it for “Sam Bradberry Taylor,” but adds, “I’m not Bret Taylor,” and most founders can’t.
After four years and meeting 5,000-plus GPs, Nel gave up on a precise GP archetype; the key lenses are “how do they make decisions and who do they hang around?” Indexing on hyper-articulate theses was “a mistake” — sometimes “this is the best founder I’ve met this year” is a back-up-the-truck signal, and the killer calibration question is “Okay, well, who else did you meet this year?” Compound’s Michael is the case study: the theses are visible, but the real edge is “a very, very good nose for who to listen to.”
Edge is cycle- and sector-dependent: in hot Bay Area frontier AI “GPs who win will win,” and if you can’t outbid Sequoia, Conviction, and Benchmark, “you probably shouldn’t play.” In colder sectors, picking wins — which is why institutional LPs were interested in Dimension, the Lux spinout focused on tech bio, as “in some sense… the only game in town.” Forward-looking opportunities include Indian life sciences, Brazil pre-seed (“so few VCs actually writing pre-seed checks”), and capital-starved China — as long as you keep the bar at “global maximum” with San Francisco benchmarks.
Concentration is “downstream of convexity,” not a rule — 40 checks at $2M post are “basically free call options,” but almost nobody besides Silicon Valley Angel has the network to do it without adverse selection. Price discipline swung from lax to overtight before settling on mentality over absolutes: what matters is getting “the best terms at the table,” because “the returns follow the mentality of the GP.” First-fund managers are backable only with a prior angel record: “if you’ve never written a check before, I have no data points to move on.”
The process lesson is constant WhatsApp debate rather than a scheduled IC. Nel’s formative scar tissue is a South African airline he bought going into COVID as deep value, only for the government counterparty not to pay and the business to go bankrupt: “I think probably there is no price.” The venture translation is emotional, not tactical — “I just don’t want to lose the money.”

Deep dive

1. Venture beta doesn’t justify the illiquidity — but the bottom-up game does

José opens with a bear case, citing Cambridge Analytica and research suggesting venture has underperformed public indices for roughly 20 years, alongside Roelof’s “return-free risk” line. Nel’s honest non-answer: “I don’t know, man, if I can answer for why people should allocate to venture broadly. I’m not sure that venture beta justifies the illiquidity premium.”
His actual frame is per-deal, not asset-class: “for every outlier company, there are at least three outlier funds. For every Coinbase, there is a Ribbit, there is a USV” — plus a dozen funds that return multiples from proximity to one company. The underwriting is “this founder in particular, why this company, why this GP, why this market.”
Pressed on whether conditions could ever kill the bottom-up case, he rejects the premise: even in 2021, “if you were fishing in Indian life sciences, probably you’d be the only one fishing in Indian life sciences.” There are always small pockets of arbitrage, and “as LPs and GPs, the job is to find that stuff.”

2. Don’t beat the big funds — front-run them and become their signal

José’s challenge: big funds have crushed it recently, their brands help with hiring and partnerships, and their platform teams may offer more than a one-person show. Nel reframes: start from the heuristic that every multibillion-dollar company will eventually have Andreessen, Sequoia, General Catalyst, Insight Partners, or a similar fund — then “you’re not trying to beat the bigger funds, you’re trying to see the companies before the bigger funds, and act as signals to the bigger funds such that they will eventually follow on from you.”
The structural reason the big funds leave room: 20-person investment teams “are not incentivized to go fishing in really strange places to find a $5 million post-money-cap founder in the backwaters.” They are incentivized to deploy $30–100 million tickets, “that’s how they drive their carry.”
His conclusion is symbiosis, not rivalry: big funds have fund-of-funds programs with pockets to seed emerging managers and use them as deal flow. “It’s not an either-or… the products are very different to LPs.” For discovery, fringe opportunities, and the signal ecosystem, pre-seed funds are where to look.

3. “Pre-consensus is different to non-consensus” — and capital sequencing is an art form

José argues smaller funds must be contrarian. Nel resists the word: you have to “bridge the non-consensus to consensus gap,” either through extreme capital efficiency, as with Hummingbird’s Turkish gaming companies, or through storytelling — a GP they both know can “take a founder and rebrand the story ever so slightly that it is suddenly a lot more palatable to Conviction, Benchmark, Sequoia.” “That ability to make somebody legible to capital is quite an art form.”
The exchange worth keeping: if a founder could raise from Sequoia or Andreessen, wouldn’t they? Nel says he would do it for “Sam Bradberry Taylor,” but “I’m not Bret Taylor,” and most founders can’t.
His counter-examples: Bridget Mendler — Disney child star, then PhD, now founder of Northwood Space — “specifically built her cap table with all these pre-seed angels” before graduating to a bigger fund. A founder from Abuja, Nigeria, “maybe one in 100 in San Francisco and absolutely 1 in Africa,” couldn’t get a term sheet for six months until Nel routed him to a risk-on pre-seed GP who could act as a signal within San Francisco hard tech.
The landing point is terminology: not contrarian but pre-consensus — “if they knew about this, and had it presented to them in such and such a light, it would be interesting to them.” It helps to know “the flavor profile of the people who will be following capital to you — not the firm, the individuals.”

4. The wunderkind pool, and the two kinds of deep tech

On Corey Levy’s fellows model, Nel calls it “total coverage of a really hot, consensus talent pool, but well before everybody else.” He places it alongside Josh Brody and the “mini Thielverse” archetype: clearly signal, though “whether it’s priced in now or not, I can’t tell.”
José’s pushback — deep-tech winners skew older — draws a useful split. Commercial deep tech includes space and government-facing companies such as Saronic, Tronomoly, and other emergent primes, where founders have practiced selling to government and have pedigreed backgrounds. Frontier deep tech includes BCIs and the very young “Neo Labs” or “Merkle” types, working at the coalface. That is “not a commercial-first business,” but “a product-first business.”

5. Bull markets are winner’s markets — so know where you can’t compete

José reads Nel’s own quote back: “When money’s worth little… GPs who win will win. When money’s worth a lot… GPs who pick will win.” Nel accepts it for what is hot: in Bay Area frontier AI, “you’re not trying to pick super well, you’re trying to win,” because talent clusters quickly, markups happen quickly, and large funds attract more talent.
The discipline: “if Sequoia and Conviction and Benchmark are trying to bid you up, it’s very hard to compete… most funds can’t find a way to compete and so they probably shouldn’t play.”
Meanwhile, colder sectors reward picking. Life science is so attractive to institutional LPs partly because Dimension — Zach, Nan, and Adam, spun out of Lux seven years earlier and focused on Series A/B tech bio — is among the few credible, proven managers able to raise and deploy capital in that cycle.

6. Picking GPs: forget the archetype — decisions and company kept

Both sides confess the same learning curve: the spiky, neurodivergent founder archetype doesn’t cleanly extrapolate to GPs, and Nomads gave up trying to create a precise one. The replacement test is “how do they make decisions and who do they hang around?” Operationally, ask, “Why did you do your last five companies?” and keep asking why until you reach gut rationale or a structured thesis.
The change of mind Nel volunteers: “we were at risk of mid-curving it for a while” by indexing on hyper-articulate theses — “I think it was a mistake.” Sometimes “‘this is the best founder I’ve met this year’ is a ton of signal and you just back up the truck,” but only if the GP’s taste is calibrated: “Have they seen greatness before? How close are they to it?” The follow-up question is, “Okay, well, who else did you meet this year?”
Compound’s Michael is the worked example. LPs can backdate the Runway ML and Wayve theses and see sophisticated inflection-timing, but “the thing people miss with Mike is who he’s hanging around with is an incredible group… a very, very good nose for who to listen to,” on both the researcher and VC sides.
Proof point: Mike and Dimension did a company together called Akira. Nel calls it “one of the best proxies for the best people currently working on that problem” among the groups trying to build “OpenAI for bio” or foundation models for life science, while preserving uncertainty about whether it will succeed.

7. The levels problem — and why you network-calibrate instead of hiring PhDs

José’s jiu-jitsu framing: he has competed and done well at an international level, yet higher-level competitors would make him look untrained. When you’re good at something, you see the levels below you and perhaps one or two above. That is why he values GPs who have “been great at something,” and worries that outside investors can pattern-match to a false founder archetype, as happened in crypto in 2020 and 2021.
Nel’s counter: for VCs, it is more a calibration-of-benchmarks game than an “I have to do it myself” game. He cites Shawn Maguire on math gradations and says you can rent calibration: “I can’t calibrate you at jiu-jitsu, but somebody else can.” Entering India or LatAm, “the first couple of things you do is just try to find a guide.”
Crucially, that does not mean automatically finding a domain PhD. “You can over-index on expertise.” Nel would rather call the Dimension guys, Demis, or other credible people and solicit deliberately opposing viewpoints — including one multi-asset tech-bio investor who believes in the approach and another who says it cannot work without putting things through the clinic.
Referencing is its own art, and competitive founder references can mislead. José describes asking a strong Indian life-science founder about a peer and getting: “He’s not as good as me — why would I put money in his company and not my own?” That is not useful calibration.
On fit, Nel invokes a quip attributed to Vinod: the transcript renders a scrappy immigrant GP as “Dagny O’Leary,” with scrappy immigrant founders such as David Veliz; Mike Moritz is associated with the more erudite Collison brothers; and Jim Goetz with the bro-y Jan Koum. “The vibe fit between GP archetype and founder matters more than the ability that the GP has in the founder’s domain.” Nel’s refinement: “I think you attract people who look like you. I don’t know if you necessarily pick them.”

8. The two-man fund: “the executor and the lazy meta thing”

Nel gives an example of two men who ultimately joined a fund together. One is referenced as incredible, universally liked, extremely early, and “of the 5,000-plus GPs that I’ve met, he’s the most charismatic, bro.” His counterpart is less conventionally charismatic — more “the thinking persona” or “the aura” — and is referenced as aggressive and transactional. “To some extent, you actually want that”; he may be difficult to share a deal with because he will move quickly.
Nel generalizes it into a strong pairing: good cop/bad cop, or the “executor and then the lazy meta thing.” “You need somebody to set the rails who’s a commercial beast, and then you need someone who’s everybody’s best friend.”
Turned on himself: “I don’t have a sob story… probably I’m the face, the pretty vibes guy, the finger guns — and my partner is the commercially aggressive bad cop. Brains, meta-thinker.” He adds that these are tropes and that impressions change with time.

9. Edge decays — back adaptation, but demand a track record

On durability, Nel is unambiguous: backing a GP “is a point-in-time thing,” and “it’s behooving on the GP that they adapt.”
Model citizen: Boris Wertz of Version One — early to SaaS in Canada, then marketplaces, crypto, India, and deep tech — entering markets bottom-up and being led by the right people. Nel thinks Boris has “a very good nose for who to listen to.”
Nel also points to Micky Malka’s riff on “fintech is dead” and his backing of the Abuja founder building drones in Africa — “the least fintech thing you could possibly think of.” The edge can be fluidity.
If you can raise a ton, your edge may not be picking at all: you can build an operating platform, use multiple strategies and asset classes, and create separate funds by vertical. For first-time funds, however, “we absolutely would, and we have — but they have to be angel investors prior.” If someone has never written a check, “I have no data points to move on.”

10. Price discipline is a mentality read; the fortunes are in chaos and fringe geographies

The pendulum confession: “in the beginning we were fairly undisciplined… then overly disciplined,” over-indexing on aggressive GPs with high ownership and low valuations. The recalibration is toward relative craft rather than absolute price: a discount, a SAFE written before the priced round, moving a month earlier, or simply getting better terms because the founders prefer you. “What matters to me is that you got the best terms at the table… the returns follow the mentality of the GP.”
José reads back his “great fortunes were made in chaos” passage, citing Rockefeller, Carnegie, Vanderbilt, the rubber barons, the Medici, the Welsers, Trump, Slim, and Dangote. He says he removed a Buffett comparison from the original wording to avoid offending Buffett.
Asked where the Medicis or smartest emerging-market investors might look now, Nel names Indian life sciences (“you’re one of two people who’s going to write checks”), Brazil pre-seed (“so few VCs actually writing pre-seed checks”), and China, which was and “still is” capital-starved. The condition: “keep the bar at the global maximum,” with San Francisco benchmarks for both GPs and founders.
Africa gets a sober treatment: limited downstream capital means capital efficiency is mandatory, and “business selection matters more than just pure-play founder selection.” Local investors may over-index on investment-banking and MBA pedigree, leaving the Thiel Fellow-type unfunded. Nel also thinks many top-tier US funds have done a good job by cherry-picking companies such as Stitch and Wave, which makes competition difficult for local VCs.

11. Concentration is downstream of convexity; debate is the process

Against treating concentration as a rule, Nel argues that “concentration is downstream of convexity” and that a portfolio can be both diversified and convex. Forty checks at $2 million post are “basically free call options,” but it is rare to find a fund with the network to do that without adverse selection; Silicon Valley Angel is the example he knows.
They agree concentration can be a symptom of conviction, selectivity, and early access. “The people you want to work with are generally very picky about their taste,” and that sort of person is often more concentrated.
The two-man process is asymmetric: Nel is intuitive and gets excited six to eight times a year; his partner is more structured and gets excited roughly three times. The cut is “super meritocratic — best argument wins,” and the partner generally trims some of Nel’s ideas.
The worked example is a GP with an appealing sector and good entry valuations who, in Nel’s view, did not understand the underlying companies and was “filibustering” around the subject. Even with a strong founder lens, “you have to understand the companies”: at minimum, how they make money.
On narrative-only investors, Nel concedes Fred Wilson telling Twitter not to monetize — “if this gets big, then you’ll figure out how to make money” — but insists that is rarer than the stories suggest. “In four years, I haven’t found an instance where a GP doesn’t have a good answer to this and is still a really good GP. Most of the time, not having an answer to it is just lazy thinking.”
Even Bitcoin is presented as an example of structured reasoning: José mentions Ribbit’s 2012 Bitcoin memo, and Nel calls it “a really structured, rational thesis” and “a really clever rationale.” José says he thinks it returned Ribbit’s fund 50x, depending on when it was sold.
José’s mirror confession is that his biggest omissions came from worrying too much about how companies would make money in nascent spaces. Nel’s reconciliation is that early crypto and frontier AI may be murky exceptions; “for most businesses, you should know.”
The process meta-lesson is constant WhatsApp back-and-forth rather than a scheduled IC with a prepared memo. The lively exchange helps partners understand how each other thinks, what questions imply, and where to push, without a rigid sequence of meetings and paperwork.

12. Value roots, fear versus curiosity, hobbits over messiahs, and the wipeout that taught counterparty risk

José’s origin story is South African value investing via r/securityanalysis. Several 5–6x public-market bets, partly beta and luck in hindsight, drew attention and became a small fund. A Tencent piece then prompted Hummingbird to reach out; they wanted to be an LP, and José reversed the pitch by asking to join them.
Value’s carryover to venture is an alpha-beta mentality: unlike the venture reflex that “if everybody likes a deal, it’s hot, so I must get into the deal,” the value lens keeps attention on places other people are not looking.
Advice to his younger self: manage the pendulum between fear and curiosity. Nel over-indexed on fear early — “fear of not having money” — and on curiosity too late. Curiosity can be positive but can also lead down bad rabbit holes; it needs to be tempered with fear.
The literary detour is a worldview statement. Against the “great-man theory of venture” and the “Nietzschean will-to-power thing,” Nel takes Tolkien over Herbert. Boromir was supposed to be the great man, but his attempt to take the ring and exercise will to power becomes his downfall. Frodo succeeds not through a messianic role or sheer strength, but through a chain of small events: he spared Gollum out of pity, and Gollum ultimately bit the ring from his finger. “Little people matter… I far prefer the world of a hobbit.”
José takes the other side: the point of Dune is wanting to be Paul Atreides, invoking Kanye’s Yeezus-era line about wanting to be Jesus Christ. Nel counters with C.S. Lewis on brave knights and heroic courage, and with Uncle Iroh’s line that power and protection are overrated and choosing love is wise.
José names Soros as an investment influence and discusses reflexivity as an alternative to Buffett-like prudence. He worries about crypto’s “crisis of confidence,” where people are mostly trying to predict what the next person will believe. Both are Taleb fans. José says Fooled by Randomness “messed me up a bit” by making him see randomness everywhere, while also finding the “you’re not a lottery ticket” framing useful.
The closer is José’s biggest mistake: he went long a South African airline into COVID as a deep-value play because the assets appeared worth multiples of the market capitalization. He failed to assess the counterparty — the South African government — which did not pay what it owed, leaving the business bankrupt. “I don’t know if I would have bought it… I think probably there is no price.” The venture application is less tactical than emotional: “I just don’t want to lose the money.”

Training a 400B Model on 2,048 Blackwell GPUs for $20M | Researcher Conversations at GTC ⁠Why OpenAI and Anthropic Won't Win the App Layer | Glean Founder