Brave startups and billion-dollar bets on AI

Knowledge base

🔏 “Welcome to the era of bold startups” by Evan Armstrong/Napkin Math: The era of the gluttonous startup is over, and we’re entering the era of determination. As venture capital funding dries up and growth slows, it’s time for a new generation of startups. Evan shares three lessons from the past (hint: stock compensation isn’t free money) and paints a picture of the future: smaller teams, bigger visions, and a focus on cash flow. Read this if you want to build a startup that can thrive in lean times.

“Explaining the New OpenAI Strawberry Model” via Dan Shipper/Chain of Thought: ChatGPT is famous for not being able to count the “r” in “strawberry.” OpenAI’s latest model, Strawberry, can solve this conundrum. It’s trained on step-by-step reasoning, which potentially makes it smarter than your average AI bear. Read this to understand why your next conversation with ChatGPT might feel more like talking to a math genius than a lucky person.

“Benevolent Billionaire or Despot – or Both?” by Evan Armstrong/Napkin Math: Properly evaluating Bill Gates requires balancing his ties to Jeffrey Epstein with his philanthropy’s saving of millions of lives—the moral math gets complicated fast. Evan reviews a new book that attempts to profile Gates. Read it for thought-provoking insight into the ethics of extreme wealth and why billionaires may represent a kind of moral oddity that falls outside our usual framework.

🎧 “Browser company builds brand that drives viral growth” via Dan Shipper/Chain of Thought: Browser Company’s Head of Storytelling, Nashilu Mouen-Makoua, reveals how she’s using AI to build a brand that doesn’t feel like a brand with the Arc browser. From giving talks in restaurants to using ChatGPT for research, Nash shares her playbook for creating authentic narratives that resonate beyond Silicon Valley. 🎧 Watch on X or YouTube, or listen on Spotify or Apple Podcasts. Subscribers get the full transcript.

🔏 “Inside the Pod: Can Artificial Intelligence Solve the Adult Friendship Problem?” by Rhea Purohit/Chain of Thought: If you feel lonely, AI can help you. New York Times columnist Kevin Roose has found 18 new friends—with AI companions. His experiment reveals the potential and pitfalls of digital friendships, whether they’re parenting advice or fitness checkups. Read on for a peek into the future of social connections and how a tech expert is using AI in his daily life and work.

Fine tuning

Alexa will soon stop sucking. Reuters reported that Amazon will swap out some of its native large language models for those developed by Anthropic. Amazon plans to offer a subscription service for $5 to $10 per month because Claude “performed better than the e-commerce giant’s own AI models.” The slowness of large companies is losing out once again to the concentrated efforts of startups. I don’t know why anyone would pay for this, but tens of millions of Alexa devices have been sold—surely some of them will pay for themselves.

OpenAI is enjoying great success. OpenAI is set to raise a new funding round at a $100 billion valuation . At $3.4 billion in full-year revenue (current month’s revenue multiplied by 12), you’re looking at a 29.4x multiple on current revenue. The unknowns here — gross margin, growth, research pipeline, etc. — really matter, but that valuation and bet would make it one of the most valuable startups in the world. Apple, Microsoft, and Nvidia are all in talks to participate, with Thrive Capital leading the round.

The new language model can absorb 750 books at a time. Magic, a startup building artificial general intelligence (AGI) with an LLM focused on coding, has announced that its new model has a context window of 100 million tokens. In terms of context (haha), most leading models have a context window of around 2 million. To accompany this research achievement, the company raised $320 million in capital. Not bad for a team of 23! I’m still thinking about what can be done with 100 million tokens compared to our current paradigm of a million tokens plus mining techniques like RAG. If you’re building something that can only be done with that many tokens, please reach out.

Nvidia’s future remains uncertain despite its excellent results. Its data center business increased revenue to $26.3 billion—2.5 times what the company generated last year. Nvidia’s total revenue and earnings beat Wall Street’s targets. Yet the stock fell about 7 percent for the week. Why? It’s unclear how long that streak will last.

The macro question in AI is: Where the hell is the revenue? On the quarterly earnings call, after an analyst asked what kind of ROI Nvidia customers are seeing on their chip purchases, CEO Jensen Huang responded, “People who invest in Nvidia infrastructure are seeing ROI right away. It’s the best investment in infrastructure ROI, compute infrastructure, that you can make today. And so one way to think about it, probably the most — the easiest way to think about it, is just to go back to first principles.”

My rule is that whenever a CEO says “first principles,” he dismisses the question. Such is the case here. Of course, every major tech company buys Nvidia GPUs in pursuit of frontier models. What’s cool is that these chips can switch back to other machine learning use cases (like ad targeting for Meta or inference for AWS) once the frontier model is trained, but those use cases only materialize as the LLMs get better and better. The AGI bounty is so compelling that I would expect this level of spending to continue, but even so, it’s not exactly a reassuring line of logic for Wall Street.Evan Armstrong

Data mining

More money…less growth? Startups that have raised less money increased revenues faster:

Source: Silicon Valley Bank.

Apart from the smallest startups, companies that raised funds for a period of less than 12 months generally recorded significant more revenue growth than those that raise funds in more than 18 months. You can’t tell much from a few bars on a graph, but one implication is that less is more when it comes to funding. Why might that be? Because if you have 12 months of runway (or less), you don’t have the time or money to be happy with yourself. Gaining a little determination could be the key to survival.—Moses Sternstein

Alignment

Prestige. Anu Atluru’s essay on “limbo” captures an experience many of us will encounter as we pursue nonlinear career trajectories. Atluru describes the limbo as a state of in-betweenness, when your professional identity becomes unclear—when you’ve left your old role behind but haven’t fully settled into your new one. In this limbo, you have difficulty explaining what you’re doing, and others may not know how to place you socially or professionally. This is especially difficult if you’ve left a prestigious career behind and taken a step into the unknown. But there’s hope: in uncertainty, you have the freedom to express yourself and explore your curiosities. Paradoxically, this period of limbo can elevate your status far beyond what it once was. Sometimes you have to get lost in order to truly find yourself.—Ashwin Sharma

Sentiment analysis

“I agree with your logic; another way to refer to these “slash roles” is to think of leadership as an expert generalizer.”—AI Founder Responds to Evan’s Article on start-ups with character

Want to chat? Send me a private message Dan Or Evan on X

Collaborative Filtering

Rule of three. Our friends at First Round Review recently published two articles that provide a useful three-part framework for founders:

Hallucination

What if Apple created a physical “Find My” tracking device?

Source: X/Lucas Crespo.

That’s all for this week! Be sure to follow Every on X on @everyone and further LinkedIn.

Sponsored by: Superfilter