Angular Ventures Weekly
Issue #173: For the week ended February 7, 2023
Subscribe here to receive new Angular blogs, data reports, and newsletters directly to your inbox.
Do ChatGPT and Dall-E 2 mark the end of an era?
Gil Dibner
One thing we get asked a lot is “what’s the impact of generative AI (ChatGPT or Dall-E 2) on VC?” There are probably a few decent possible answers to this, but here’s one take I had recently:
Generative AI is — at least right now — mostly a proof point on how advanced software/data/ML/NLP has become. The easy UX that ChatGPT and Dall-E 2 offer have made it really easy for people to understand how far AI has come. We are all playing with these new toys which is driving that point home query after query.
It’s important to remember that the quality bar for generative AI is actually lower than for classifiers or predictive engines. There is little business impact if the image of a squirrel surfing is not a certain way. But this is not true for classifiers (is this fraud?) or predictive engines (who will churn? will this machine break?), where the business impact can be bigger — and the data more proprietary. This matters because these are still, fundamentally, toys and not proper enterprise-grade tools. The generative AI tools will come soon enough, but they are not here now.
But I digress. The impact of the popularity of the ChatGPT/Dall-E 2 experience is mostly about a realization — one that is more powerful for people who are less connected to actual software applications. This new tech has made it pretty easy for everyone to see just how far we’ve come — how powerful the modern data stack and modern AI models can be when applied to the right data. At Angular Ventures, we have been operating under the assumption that AI/ML/DL tech is on the way to commoditization for some time. I wrote a post back in 2017 that outlines this thesis: “We are in the midst of a “technology value inversion,” in which a lot of the technologies that once created value for new entrants (like SaaS or AI) will potentially destroy value for new(er) entrants because they have been fully adopted and commoditized.”
The real impact of the popularity of ChatGPT/Dall-E 2 is that they are going to suck the perceived value out from under the feet of hundreds of startups that were trying to monetize on capabilities that are now pretty easy for anyone to implement. As we all get to play with these previously magical toys, the magic starts to be exposed for what it is, and it will inevitably start to seem less and less magical. This feels a bit more like a period ending than a period starting. We’ve taken the most powerful technology we have and turned it into a parlor trick…because it’s finally cheap and easy enough to do so.
I’m still thinking through the implications of this…but in some sense — for us at least — nothing has really changed. We’ve always believed that it’s very hard to create real enduring enterprise value in software even if it’s built on a bedrock of AI/ML/DL etc.
As a venture investor, I always take this approach: either (1) show me something no one knows or (2) show me something no one else can do. The superset of those two things is the space relevant for VC investment. We have always known that space is very small — and to me it feels like a lot more people are going to realize this over the next few months.
If you are building based on something no one knows or something no one can do, reach out. We want to hear about it.
Gil
EVENTS
Feb 8 / Lessons Learned From Investing Early in Over a Dozen SaaS Unicorns Including Salesforce, SuccessFactors, Box, Gusto, SalesLoft, ServiceMax, Veeva, Bill.com, Doximity, Yammer and Zoom Among Others
Jason Green, Founder & General Partner, Emergence Capital
Mar 8 / The Evolution of Collibra’s Product Positioning & How They Created a Category
Stan Christiaens, Co-Founder & Chief Data Citizen, Collibra
FROM THE BLOG
The Hierarchy Trap
Hierarchy is an important tool for providing structure and alignment, but it can easily grow like a weed if not managed.
The Tech Recession of 2022
What lessons can we learn from the successes and failures of the 2010s?
No One is Having Fun Right Now
And that’s okay.
It’s Never too Early to Build your Growth Model
What are the specific mechanisms by which one user turns into many, and an initial investment turns into revenue?
EUROPE & ISRAEL FUNDING NEWS
Spain/SME ERP. Jeff raised $90M for its business in a box for entrepreneurs with a platform that offers web, app, and physical PoS options.
France/Space Systems. Exploration closed $44M for its reusable orbital vehicles.
UK/Climate. Risilence raised $26M for its enterprise climate risk management software.
UK/Industrial Systems. Recycleye raised $17M for its automated waste sorting systems.
Spain/Design. Penpot raised $12M for its open-source software for design and prototyping.
WORTH READING
ENTERPRISE/TECH NEWS
Breaking the AI bottleneck. Crunchbase has a piece this week called ‘AI In 2023: A Year of Reckoning In The Enterprise’. In a bid to roll out these new advancements in the models, enterprises are hitting a familiar problem. “Almost every CTO I talked to in 2022 said their biggest roadblock to deploying AI was finding the right engineers and data scientists to help them get there.”
Incumbents roll up AI features. Microsoft just launched Teams premium, powered by ChatGPT, at just $7/month. With ChatGPT, Teams users can generate automatic meeting notes, AI-recommended tasks, personalized meeting templates, and a lot more. Intercom announced a wave of new features using ChatGPT to help their user base with customer service with the summarization of tickets, writing responses, and rewriting text to find the company tone.
Taking a different angle to solve AGI. A Q&A with John Carmack on the different path that he is taking to solve Artificial General Intelligence. “The reason I’m staying independent is that there is this really surprising ‘groupthink’ going on with all the major players. It’s been almost bizarre in the last year to see things like: OpenAI releases an image generator, then Google releases one, then Facebook releases one. So, all of these companies are just within a couple of months of being able to reproduce anybody else’s work, because they all draw from the same academic researcher pool. There’s cross-pollination and an enormous brain trust of super-brilliant people doing all this… But, because we do not have that line of sight — we’re not sure that we’re in the local attractor basin where we can just gradient descent down to the solution for this — it’s important to have some people testing other parts of the solution space as well.”
ChatGPT competitors. Claude, which is Anthropic’s competitor to ChatGPT, is out. It sounds like Google will unveil its ChatGPT clone this week on Wednesday, February 8th.
HOW TO STARTUP
Incentives drive outcomes. Sam Blond, former CRO at Brex, talks about how they made one change at Brex to 3x SDRs productivity.
Early signs of PMF. Miles Grimshaw of Benchmark quote tweeted an old Jeff Bezos clip where he relays a story of a buyer from Bulgaria who mailed a floppy disk with money stashed inside the metal box to pay Amazon. Miles remarked, “one of my favorite anecdotal stories of early PM fit. A story of determined customer pull in the first 6 months!”
HOW TO VENTURE
Mass startup extinction on the horizon. Tom Loverro from IVP’s Twitter thread has gone viral this week. “PREDICTION: There’s a mass extinction event coming for early & mid-stage companies. Late ’23 & ’24 will make the ’08 financial crisis look quaint for startups. Below I explain when, why & how it will start & offer *detailed advice to founders* on surviving the looming die-off.”
PORTFOLIO NEWS
Reco’s CEO, Ofer Klein, was recently profiled in Authority Magazine. “As a business owner, you have gut feelings about various aspects of the business, partners, employees, customers, etc. Knowing that nine times out of 10, your gut feeling is correct. All I can say is trust yourself, your mentors, and ultimately, stay the course to get where you want to be.”
Datos Health and Shirley Ryan AbilityLab collaborated to study the effectiveness of remote therapeutic monitoring on patients with parkinsons disease, multiple sclerosis, and those recovering from stroke, brain injury, and more.
