A small ask: I’m on a mission to get Built By Berkeley to 10,000+ subscribers this year. If you enjoy the newsletter, please pass it along to someone else who might
Deal of the Week
Always nice when a deal I invested in makes the deal of the week. And even more fun when I know there is a bit of a backlog of them to come out over the next few weeks.
This week it is SweetAg (previously Landjourney, where I invested in the pre-seed), which just announced a $7.4M round to accelerate its AI-native lending platform for farm lenders.
Farm lending is a massive use case for AI, with a few different quirks—land parcels, livestock, etc.—compared with more traditional lending that SweetAg helps solve. AI can supercharge this process by onboarding a ton of information across different formats.
SweetAg has already processed ~$10B of originations through the platform, so it is firmly off to the races. Onwards!
Congrats to Luke Johnson (Haas MBA) and team!
The Pelican Index
As I constantly interact with AI on a daily basis, I think you sometimes need to step back and appreciate the rate of change. How often do you now experience hallucinations? How often is the Excel wrong or the command misunderstood?
A great blogger, Simon Willison, has what he calls the Pelican Index, where he asks models: Generate an SVG of a pelican riding a bicycle.
If you go all the way back to the proverbial dark ages of October 2024, you got this:
Now it looks like this (full quality):
It might be a slightly silly way to measure progress, but it does feel fairly remarkable that in such a short time these models have just got really, really good.
Jensen Huang of Nvidia recently declared that AGI has arrived (conveniently trained on some of his GPUs 😉).
I have spoken previously about the ‘end of busy work’. I think the models are finally here that will deliver that, and I have felt it especially over the last few weeks in the finance domain. Fable within Excel is now like vibe coding—or maybe vibe financing!?
It really is amazing. One of my brothers, who works in London and was previously an AI skeptic, is now asking me to run things through Fable for him (he cannot get access in the UK!).
Slowly at first, then all at once.
The one that got away….is still getting away.
A while ago I posted about some lessons. One of those was:
𝗟𝗲𝘀𝘀𝗼𝗻 𝗧𝘄𝗼: 𝗚𝗿𝗲𝗮𝘁 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗮𝗿𝗲 𝗮𝗹𝘄𝗮𝘆𝘀 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲.
About two years ago, I tried to invest in a company. The CEO said the round had closed, but they'd let us in, at a 25% premium to the just-closed round. Felt a bit off. I politely declined.
Today? I hear they are about to raise at a $500M valuation and are growing exponentially…so if I had been able to invest at $19M valuation vs. the $15M I would have still been a happy camper with my 26x increase.
Well that company is now at over $500M in revenue and looking like it will be more like a $4-5B valuation soon. So now more like a 263x 🙂
All I can really hope is that by the time I end my career as a venture capitalist, I have many, many more of these (and maybe a few big wins!).
Because if you are not seeing these types of companies, are you really in the flow of great deals?
Is the personal AI agent going to be the breakout consumer use case?
Silicon Valley is abuzz with personal AI agents from Grokbot, Town and now Instinct.
I have been playing around with Instinct (recently valued at $2.5B pre-revenue) and it is fairly magical. The UI is simply iMessage or WhatsApp, and you can ask it to do things—fill in forms, research, book things, etc.—in a way that feels very natural.
It feels like a great consumer form factor: conversational, always available and, in theory, increasingly able to learn what you want over time.
There are still a lot of questions to answer:
Privacy: Instinct asks for fairly broad access to your data, which will understandably make some people nervous.
Guardrails: When does the AI need guardrails? It is early, so I think this will get ironed out, but there are already examples of personal AIs changing passwords, deleting accounts and generally taking actions in pursuit of an outcome that the user may not have explicitly intended.
Will booking change forever: What happens to things like restaurant reservations in a world where everyone has an AI continually checking availability? One Instinct user got banned from Resy after his agent kept hitting the API trying to get a spot at a hot NYC steak restaurant. For top restaurants, maybe they eventually start charging for reservations!?
What does this mean for ecommerce if this takes off? Agents will optimise for price, reliability and speed rather than brand or browsing experience. Does that strengthen platforms like Amazon and Shopify. Or does the agent itself become the new distribution layer?
If this form factor takes off, the world is going to look and monetise very differently.
My Meta prediction still looks either early (or wrong)…but surely they should be well placed to win this market, or at least buy one of these whippersnappers for some $20B or so.
Meta is expected to launch its own version of an AI agent. Hatch in the coming weeks.
Top of the Pops
The annual Pitchbook startup database is our and Berkeley is top again for under graduate founders, number 4 for graduate, and 9 for MBA
So a great achievement for undergrads and more work to do for the rest.
Interestingly Berkeley only took the top spot a couple of years ago post the starting of Built By Berkeley and I know we have unearthed a fair few founders previously not captured by Pitchbook.
Quick Takes
We have spoken about Recursive Learning as the potential next big breakthrough. To me it seems we are getting pretty close / may already be there in the big labs. Check this article out from an OpenAI researcher. If the AI can start to train itself the implications become pretty large…
Summary by the #️⃣ & 💰:
4 Berkeley-founded companies funded
$16M of capital raised from the 31st August to 6th September
💡 Got any ideas or feedback on how to improve this weekly digest? Just hit reply.
Acquisitions
🐻 Jen Zeszut, CEO & Co-Founder. BA Psychology Article
🐻 Aaron Sokol, Co-Founder & CRO. Haas MBA Article
Closed Rounds
🏡 Ollie AI. $7.5M Seed 🇺🇸 AI assistant for families. 💰 Khosla Ventures, AI2 Incubator, ADS Ventures
🐻 Bill Lennon, Co-Founder & CEO. BA Maths Article
🌾 SweetAg. $7.4M Seed 🇺🇸 AI platform for agricultural lending. 💰 Builders VC, Courtyard Ventures, Diagram Ventures
🐻 Luke Johnson, Founder & CEO. Haas MBA Article
🧬 GonGlobal. $1.2M 🇦🇺 Needle-free IVF drug delivery. 💰 Flying Fox Ventures, LaunchVic, UNSW Founders
🐻 Ellen Gonda, Co-Founder & CEO. BA German & Communications Article
📓 Clusy. Undisclosed 🇺🇸 AI-native IDE for Jupyter. 💰 Founders, Inc.
🐻 Eldar Hasanov, Co-Founder & CTO. BA Maths Article
Date Built By Berkeley Started | Companies Funded | Total Raised ($M) |
7/8/24 | 852 | 234,023 |
Our goal is to document the startup ecosystem of Berkeley-founded companies. Please share this newsletter with any Cal Bears in your network so we can crowdsource information about all investment rounds and job opportunities.
Did we miss a company or want to announce a round? Add it here, and we will post it next week.
Do you have a job you want to post from a Berkeley Company? Add here.
Is there someone in the Berkeley ecosystem that you would want us to do a profile on? A founder, funder, or general startup person? Add here.
Built By Berkeley is not affiliated with UC Berkeley, but maybe we will be one day if we get enough subscribers….
Built By Berkeley, where we announce all the funding rounds by Berkeley-founded companies. This is a community effort, so please let us know if we missed a company here. 🐻
