Nov 12, 2026 · 1h 21m
Listen from 11:40
Listen at 11:40
AI Rewrites Venture Capital’s Playbook
Founders Are Now Sending VCs a File Written for AI
As AI changes how startups are built, evaluated, priced, and financed, investors are reassessing which advantages remain distinctly human.
- 1AI-readable fundraising files may become a new signal as founders optimize materials for machine analysis and agentic discovery.
- 2Vibe-coded software lowers creation costs, but distribution, proprietary data, security, compliance, and specialized communities remain meaningful moats.
- 3SPVs are drawing family-office capital toward direct deals while widening the divide between mega-funds, early-stage investors, and the struggling middle.
Don't miss
D.A. Wallach challenges the idea that AI will cure cancer, Alzheimer’s, heart disease, and mental illness within a decade, forcing a debate over evidence and false hope.
The brief
D.A. Wallach, Jenny Fielding, and Ben Lerer examine a venture market where AI is changing fundraising materials, investment analysis, software economics, and capital flows.
Founders are beginning to send Markdown files designed for large language models, raising a new question: does AI readability become a fundraising signal or merely another optimization game?
The panel separates AI-assisted analysis from venture capital’s human work: building trust, challenging founders, supporting companies through crises, and exercising judgment when evidence is thin.
Jelly, Everywhere Ventures’ research platform, reflects the shift from expensive data products toward tools that help investors investigate companies, investors, and market conversations.
The discussion widens to vibe-coded software, SPVs, and medical hype, arguing that cheaper creation and easier access do not erase distribution, diligence, governance, or ethical responsibility.
The sharpest warning targets sweeping claims that AI will cure major diseases within a decade, putting probability, evidence, and the consequences of false hope at the center.
What was said on this episode
37 statements · 21 positive · 9 negative · 1 mixed · 6 neutral
Everywhere Ventures invests at pre-seed across health and work and has backed 300 companies.
“early stage pre-seed fund, investing across many health and work. And we've invested in 300 companies, so pretty active.”
Listen at 1:21
At least one-third of Everywhere Ventures’ biggest portfolio companies have pivoted.
“I'd say in our portfolio, at least a third in some of our biggest companies have pivoted.”
Listen at 4:52
Technology’s current upheaval will produce more startup pivots.
“you may see even more of it now because the ground is just moving beneath our feet in ways that it. only does during a real sort of reshuffling and, you know, revolutionary moment in technology.”
Listen at 5:06
Software startups can test a direction for six months and then pivot.
“Whereas software companies with the pace of development right now, you can do something for six months, decide it's not the thing”
Listen at 5:43
Biotech companies are largely locked into an indication through their first human trial.
“once you've selected the indication or the disease you want to go after, you're kind of locked in at least through that first human trial.”
Listen at 6:27
About 99% of reviewed VC websites were not optimized for agentic discovery.
“almost no, it was like 99% of the websites that we looked at of other VC firms weren't optimized.”
Listen at 9:47
AI-readable fundraising materials represent the next evolution of agentic venture.
“I think this is the next evolution of, you know, agentic venture and I'm all for it.”
Listen at 10:36
AI-readable company files signal willingness to be evaluated on genuine merits.
“you're going to be a company that is going to, you're comfortable trading on your merits”
Listen at 12:42
AI provides high-quality analysis and market mapping for venture investing.
“we get really great analysis and good market mapping and good work.”
Listen at 14:44
AI venture recommendations are consistently low conviction.
“the recommendations from AI are. no matter what we do, really disappointingly low conviction.”
Listen at 14:51
Their AI scoring system produces scores of three, four, six, or seven nearly every time.
“we go and we put it in and we get scores. between three, four, six, and seven, almost 100% of the time”
Listen at 15:21
Algorithms may become better at selecting venture investments than humans.
“I actually embrace the idea that the algorithm is a better picker.”
Listen at 23:41
The venture capitalist profession will not continue in its current form.
“I don't think that this job will really exist.”
Listen at 24:22
Everywhere Ventures is open-sourcing and potentially selling Jelly to other managers.
“we've started open sourcing it and actually, you know, providing it to other managers and potentially selling it.”
Listen at 29:41
Everywhere Ventures plans to sell Jelly soon.
“we will be, you know, selling this product soon.”
Listen at 30:25
Vibe-coded products without distribution or data advantages lack durable competitive strength.
“without a distribution or a data advantage, you know, you're really just vibe coding.”
Listen at 33:22
Enterprise-ready software requires hardening, security, and SOC 2 compliance.
“building it so that it can really truly support the sort of nuanced needs and the security concerns and the SOC 2 compliance”
Listen at 33:30
Instinct’s user pull has contributed to a $30 billion valuation.
“instinct has gotten into the zeitgeist and has the user pull. And as a result is now worth $30 billion.”
Listen at 34:31
Slack continues performing exceptionally well and was likely Salesforce’s best acquisition.
“Slack continues to perform exquisitely well is probably the best buy that Salesforce has ever made.”
Listen at 35:00
A broad SaaS collapse has not yet occurred.
“the Sasspocalypse of it all hasn't happened yet.”
Listen at 35:07
Black Letter charges clients for completed legal work rather than time.
“It's an AI first law firm where you pay for the work, not for the time.”
Listen at 38:08
Venture investors increasingly favor younger, faster, more AI-native startup teams.
“we are favoring teams that we think are gonna move faster and faster and faster and faster and be more and more AI pilled”
Listen at 39:30
Family offices allocated 34% of VC commitments to SPVs, up from 9% in 2019.
“family offices put 34 percent of VC commitments into SPVs, up from 9 percent in 2019.”
Listen at 40:23
Family-office investment through SPVs is a genuine trend.
“I think this is a real trend.”
Listen at 43:03
The number of emerging managers closing funds fell from 927 in 2022 to 211 through Q3 2026.
“in 2022, there were 927. In 2025, there were 448. And then 2026 through Q3, 211.”
Listen at 45:51
Late-stage private SPV deals should be compared with public-market investments.
“When you're looking at a $8 billion private company raising a Series C through an SPV, you should be comparing that to public. comps.”
Listen at 49:54
Late-stage SPV deals have different valuations, outcomes, and risk-reward than early-stage venture funds.
“The dispersion of the outcomes is going to be radically different. The valuations are radically different and the risk reward is very different.”
Listen at 50:10
The middle segment of the venture market is severely impaired.
“The middle is cooked.”
Listen at 53:42
SPV activity will generate many lawsuits in the next one or two years.
“a lot of lawsuits. A lot of lawsuits.”
Listen at 57:29
AI researchers should express a 5% probability of curing all diseases within ten years.
“why can't you just say, we think there's a 5% chance that we're gonna cure all diseases in the next 10 years.”
Listen at 1:08:48
A utopian AI outcome has extraordinarily low probability.
“I think there's utopia, P utopia, extraordinarily low likelihood of this.”
Listen at 1:10:24
All diseases being cured within ten years has a 1–5% probability.
“I think my P. Dario is between one and 5%.”
Listen at 1:12:01
Phantom Neuro developed an implantable device placed beneath skin and above muscle.
“what they've developed is an implantable device that can go right under a patient's skin and above their muscle.”
Listen at 1:14:11
Phantom Neuro’s implant enables thought-based control of robotic prosthetic limbs.
“you can control a robotic prosthetic limb just by thinking about it.”
Listen at 1:14:37
LaunchPoint HQ offers scalable, high-performing user-generated-content advertising.
“there is a new channel to spend money at scale that is performant like Facebook or Google or Insta.”
Listen at 1:17:29
Obliteration AI lets vetted cybersecurity teams use uncensored models such as GLM-53.
“you can use obliteration AI to use things like GLM-53”
Listen at 1:19:57
Obliteration AI helps cybersecurity teams identify vulnerabilities using unguarded models.
“the ability to see what these models can do without the guardrails on them in order to find all the leaks in their cyber security.”
Listen at 1:20:16
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.