a16z

Where Startups Can Win Beyond OpenAI and Anthropic | HubSpot’s Dharmesh Shah: summary

YouTube summary9 sectionsWatch on YouTube ↗

This is an AI-generated summary of the YouTube video "Where Startups Can Win Beyond OpenAI and Anthropic | HubSpot’s Dharmesh Shah" (a16z), made with Samuraize and published by Samuraize. It condenses the YouTube video into 9 titled sections you can read in a couple of minutes, each linking to the moment in the video it covers.

1
Filed under📈 Business0 comments🍱 Add to trayReport
Study this

Where Startups Can Win Beyond OpenAI and Anthropic | HubSpot’s Dharmesh Shah

a16z

切

AI as a new layer 0:00

HubSpot sees AI as another shift in software abstraction, like the move from machine language to C and then higher-level tools. The point is not that business problems changed. The point is that you can now solve them with a better tool set and build at a higher level. That also opens the door to more people becoming builders, because the tools in the chest are larger than before.

切

Faster building, harder distribution 6:00

AI makes it easier to test ideas quickly, so you can build something, put it in front of users, and learn fast instead of waiting months or years. But distribution is still hard, and maybe harder, because more people can build software now. That means more noise and more competition for customer attention. The answer is to start narrow, focus on a small first use case, and shorten the learning loop as much as possible.

切

Two product lessons 13:01

HubSpot first tested a narrow ERP idea for law firms, then learned that the real pull was marketing and getting more customers. That led to a marketing product for small businesses. It later became clear that many users needed a CRM but found it too hard and too expensive, so HubSpot made it free and simple, then grew it into a platform by solving the whole customer problem.

切

Ecosystem and AI risk 21:00

On the ecosystem side, the goal was to be open and friendly, unlike platforms that shut developers out after attracting them. HubSpot tries to build trust by being clear about what it will not do and by helping developers reach value quickly, from first API success to first dollar. On AI, the danger is dismissing OpenAI and Anthropic too quickly. If “not yet” still fits, or if the company can do it, you should be careful.

切

Model companies and distance 26:01

Stay a little off the road the frontier model companies are widening. Choose markets they are unlikely to enter because of regulation or because their own incentives make it improbable. A company like OpenAI could chase HubSpot’s CRM revenue, but the very ambition that makes it capable also makes that move unlikely. The safer bets sit off the yellow brick road, where large model makers are focused on selling inference and tokens.

切

Interfaces and primitives 32:01

ChatGPT and Claude can become new browsers for business software, but the interface is not always where the value lives. In many B2B products, the work behind the screen matters more than the front end. HubSpot is built on reusable primitives, so when the company added sales and CRM, it could remix the same core pieces instead of starting over. That kind of structure lets you adapt as models and markets change.

切

Context and memory 39:02

HubSpot is treating AI as a new base layer for the next 20 years. The company still needs a structured store for known facts, but it also needs a system of context for unstructured data. That context may look like a graph. Used together, the two can feed inference, agents, and work done. Agentic coding also lets the team revisit old choices, replay code paths, and test them with evals, which replace older unit tests.

切

Startup inside HubSpot 43:01

To move fast, HubSpot runs HubSpot Next as a startup within the company. Uspot is one current project, a one-person CRM for solopreneurs. It is used to test a second-brain idea, with social data and context pulled together for one user. Some projects stay standalone, some get absorbed into HubSpot, and some ideas are folded into core products, much like an aquihire without buying anything.

切

AI expands the market 52:31

He says the market for small businesses is still large enough, even if it has always been hard to sell into. He compares AI to software, not a new force, and says it opens a bigger set of problems to solve. Some old software businesses will fade, but the overall chance is larger than before. If you build near software, this is a very good time.

AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.

Summarize your own YouTube video

Paste a YouTube link, article, PDF, ebook or slide deck and get a summary like this in seconds. Free to try, no sign-up needed.

⚔️ Try the YouTube summarizer

Discussion

Sign in to join the discussion. Sign in

More from the Bento Box

Browse the Bento Box →

With your OK, Microsoft Clarity and Google Analytics set cookies (data goes to the US) so we can see what breaks. Cookie details