The psychological boomerang of generative AI
Gil Dibner

The observation that today’s startup fundraising landscape divides into two camps is now commonplace. While the majority of startups are trudging through the badlands of cost cutting and challenging fundraising, startups that have latched onto the generative AI boomlet (or talk enough about LLMs or vector databases) seem to have endless capital available at whatever terms they want. What has caused this sudden exuberance?

I’m not convinced the excitement is entirely justified. I have no doubt that generative AI (or, more accurately, LLMs and their ability to process unstructured natural language data) will create value. The newly apparent capabilities of these technologies are undoubtedly impressive, but the readiness of these solutions for prime-time is far from certain. Furthermore, I wonder about the ease with which a startup can penetrate this market and build a sustainable, defensible, and sticky business model. This worry is magnified by the proliferation of foundational models (and, soon, capabilities such as voice synthesis) developed and deployed by the tech giants. The emergence of powerful open-source alternatives with increasing capabilities also calls into question many of the business models we are seeing in our dealflow. Every day brings several new opportunities for us to invest in a “generative SQL platform” or yet another “LLM-powered copilot for marketing/success/product/engineering/security/etc….” When we see two identical companies emerge at the same time, it’s an indication that something is happening. When we see two or more a day for weeks on end, it’s a sign that founders are racing to ride a hype curve, but perhaps without a lot of staying power. Will LLMs and GenAI be transformational? Yes. Will some startups achieve massive success? Of course. Is the hype that is driving up valuations and round sizes justified? Probably not.

So why are VCs playing along with this? My theory on this is mostly psychological. The troubles of VCs right now are well documented. Even Sequoia is shedding partners and entire regions, and most VCs are not doing much better. Markdowns and write-offs are starting to hammer VC returns, and we all know there is more ahead. To my mind, the unbridled VC hype cycle we are witnessing right now is a natural emotional response to the threats VCs see all around them — and, specifically, within their own portfolios. The rapid pace of tech evolution coupled with the crashing market is threatening existing portfolio companies more than it is creating compelling opportunities for new investments, but it’s emotionally easier to focus on the promise of the new rather than struggling to contain the damage from the old. Most people will believe what they want to believe, and the idea that generative AI is so exciting that it justifies throwing caution to the wind and shoveling money at seed-stage opportunities as if they are post-PMF Series B companies is what VCs want to believe. It is what they need to believe to avoid taking a hard enough look at their existing portfolio or their reckless behavior over the past few years.

VCs and many founders desperately want to believe in a new hype cycle, but wanting something to be true doesn’t make it true. That is the primary lesson of the last few years.

PODCAST

US Immigration Best Practices
Jennifer Schear, Founding Partner, Schear Immigration Law Firm

Learnings from Co-Founding Peakon, Podio, and Future Five
Kasper Hulthin, Co-Founder, Peakon, Podio, Future Five

The Evolution of Collibra’s Product Positioning & How They Created a Category
Stijn “Stan” Christiaens, Co-Founder & Chief Data Citizen, Collibra

Lessons Learned From Investing Early in Over a Dozen SaaS Unicorns
Jason Green, Founder & General Partner, Emergence Capital

FROM THE BLOG

Kafka Survivors of the World, Unite!
Why we backed Memphis.dev.

The Problem with Startup Advice
The best founders and investors know the rules, but also know when to break them.

Looking Back to Move Forward
How to survive this extraordinarily exciting and wildly disconcerting age of generative AI.

LLMs and the Future of Customer-built Software Design
How will LLMs change software development and design?

WORTH READING

ENTERPRISE/TECH NEWS

Llama 2. Meta announced Llama 2, their new LLM and competitor to ChatGPT. Llama 2 is open source and free for both research and commercial use, unlike Google and OpenAI’s models. While it is open source, companies with over 700M MAU will require Meta’s permission to use it. Additionally, it does not have a simple consumer-friendly interface like ChatGPT does, “as it is aimed toward more technical and corporate use cases.”

AI designed chips. Google’s DeepMind is working on designing faster and more efficient chips with AI. “DeepMind’s AI-based approach, which it began working on about 18 months ago, focuses on making improvements to logic synthesis, a chip-design phase that involves turning a description of a circuit’s behavior into the actual circuit. Computer chips are made up of millions of logic circuits or “building blocks,” said Sergio Guadarrama, a DeepMind senior staff software engineer. While it is easy to optimize a few of them manually, it is impossible to tackle millions of them, he said. By applying AI to speed up the design of logic circuits, DeepMind’s goal is to make the design of specialized chips more automated, efficient and less reliant solely on the work of human hardware engineers. That is a difference of thousands of designs generated by AI in one week, compared with one design produced by a human in a few weeks, Guadarrama said.”

HOW TO STARTUP

Dark skies. As Gil wrote a few weeks ago, it’s a tough time to be a founder. In light of layoffs, evading PMF, muted valuations, slowing sales, and frozen IPOs, many startup founders are now forced to operate in the dark. Tomasz Tunguz muses that “this is the part of the startup cycle that’s the grit-your-teeth grind. Some startups focus on product development. Others focus on winning competitive share by investing in sales & marketing. Having that focus enables companies to continue to execute through the dark skies & maintain a level of urgency in the organization throughout the storm to be well positioned once the rain abates.”

Venture-backed vs bootstrapped growth. New research from Capchase found that early stage SaaS startups grow at similar rates regardless of venture funding, in fact for “SaaS startups with between $1 million and $15 million of ARR saw nearly identical levels of growth, on average, over the last year regardless of whether they raised venture capital.” Both startup cohorts grew by a little less than 45% YoY. One would expect growth to be higher for venture backed companies, so why isn’t that the case? Miguel Fernandez, the CEO of Capchase, thinks that the surprising results are because “bootstrapped businesses have been a little immune to the downturn in tech, and the reason why is that they live in a perpetual downturn; there is no money for them. They have to make sure their spend is very aligned to revenue and cash flow.”

HOW TO VENTURE

The partnership. Last year, Fred Wilson wrote up a blog post on the ideal partnership structure for early stage investing that is worth revisiting. Fred shared that in his opinion, “the small flat partnership is the best structure if the goal is to produce high return on capital funds.” He concludes that for VCs “the key is partnership. Real partnership. A real partnership is where everyone is equal, not just in terms of economics (which is critical to sustaining this model), but also in terms of influence and stature. This is actually quite rare in the venture capital business. I see it in some other firms. But I don’t see it very frequently. The firms that have this are special places. They are special places to work at. And special partners to take capital from.”

PORTFOLIO NEWS

Aquant’s CEO, Shahar Chen, shared an overview of vertical AI, the next revolution in GenerativeAI.

Steadybit CEO, Benjamin Wilms, was featured in TechCrunch where he discussed the benefits of conducting chaos testing in development.

Forter leaders, Dany Naigeboren and Galit Shani-Michelwere, were appointed to the Merchant Risk Council.