Tools & Platforms in 2023
Great startups have a good answer to the “why now?” question. What is changing about the world that allows for a new company (often going against the grain) to grow rapidly? Sometimes, macroeconomic (ZIRP) or regulatory (Pix) forces act as catalysts for change. But many times, there is some technological toy or primitive that a smart entrepreneur can build with or on, remix, mold, deploy to, etc. forming the foundation for new companies. As an example, one thesis is that new iOS APIs provide developers with scaffolding to build global & open social graphs on a default local & closed Apple ecosystem – recent examples include BeReal with multi-camera capture and Locket with WidgetKit. Both startups and big companies can release these types of building blocks.
I’m highlighting where specifically I’ve seen pockets of entrepreneurial energy and what feels exciting amidst a backdrop of macro doom and gloom. For the purposes of this post, I separate tooling (used within a development workflow, acting as an ingredient, replacing repetitive processes) and platforms (unique data, ecosystem, or distribution that startups then leverage). AWS is a canonical example of the former, and the iPhone / iOS the latter. The following list is more observational, not comprehensive and the categorization isn’t perfect either. I’d also point you to my friend Alex Mackenzie’s work - he does a great job with technical deep dives on these topics.
Tooling
These are tools I’ve seen out in the wild recently or have come up frequently in conversation with entrepreneurs. It’s also important to note that not all of these technologies can support independent commercial entities. For example, programming languages themselves are notoriously difficult to monetize.
Bun - New Javascript runtime designed for speed.
Drifting in Space - Building infrastructure for the next generation of browser-based applications. Enabling live, collaborative, real-time experiences.
DuckDB - In-process and serverless analytics database. Can be used in new data engineering workflows or allowing for richer, data-intensive end user applications.
Dust - Development and deployment tool to build applications on top of LLMs.
eBPF - Kernel programming technology. Serving as the underpinning for new companies in networking, security, and observability.
Hugging Face (Transformers) - APIs and libraries to download, train, and use pre-trained NLP models. Clear frontrunner to own the “GitHub for NLP / ML” opportunity.
LangChain - Library designed to help developers work with and extend LLMs (model-agnostic) in their applications. Search Twitter for extensive examples of chatbots and vertical search products people are building.
Nix - Build and package management developer tools. Picking up popularity in companies with complex computing / hardware environments.
OpenAI (DALL-E, GPT, Codex) - Somewhat self explanatory these days but OpenAI is at the forefront of large language model research and development. Their ambitions are broad across text, code, image, video, 3D, etc. May also fit into the platform bucket.
Pinecone - Vector search management and database. Powering new semantic search applications.
Precise Point Positioning - Navigation technique for locations with accuracy errors under centimeters. Used for IoT edge deployments, self-driving / autonomy. API providers include Swift Navigation and Point One Navigation.
Replicate - ML deployment API. Often being used with Vercel for ML-powered web experiences.
Retool - Internal tools builder, but have also heard anecdotally from founders in operationally complex businesses like fintech and logistics that Retool helped get their startup off the ground. Also has platform potential if devs can build “Retool native” apps.
Ruff - Extremely fast Python linter and code transformation toolchain. Some of the most popular Python projects (like pandas and Transformers) have adopted Ruff.
Stable Diffusion - Open source large text-to-image model. Adopted by artists, developers, and hobbyists for visual content. Apple has already released support for CoreML.
Starlink - High-speed, low latency broadband across the globe. Can help power connectivity for vertical software or robotics in the terrains, oceans, and the skies.
Stedi - API for EDI integrations. Helps startups in regulated industries or those who work actively with large legacy companies. An example here.
Supabase - Open source Firebase alternative. Incredibly popular for building new projects and known for developer experience.
Vercel - Originally the commercial entity supporting Next.js, now has built other tools for frontend and web teams. Used by both hobbyists and large enterprise.
WebAssembly - Emerging as a substitute for containers in both client-side and server-side applications due to performance, safety, and compatibility.
Related: WebGPU
Zig - General purpose programming language designed as a drop in or replacement for C / C++. TigerBeetle is written in Zig.
There are plenty I’ve missed – most notably in gaming, healthcare, crypto, robotics, and biology. Would love to hear of examples if any come to mind. And of course, there are things that aren’t new but continue to see development usage and growth – Stripe, Plaid, Ethereum, etc.
Platforms
The number of new platforms feels somewhat limited, but if something does emerge as one, it is a very big deal. Startups leverage platform’s internal capabilities in some way (perhaps a unique dataset) to enhance their product experience as well as distribution (some captive audience or customer base they can direct attention to). Some have also mentioned ETH L2s like Optimism and Arbitrum, but it’s a space I know less well.
Figma - Well known design tool expanding its push for plugins that augment and extend the Figma experience. One natural starting point for AI apps.
Grafana - Ubiquitous dashboarding solution for metrics and observability. Has an integration-friendly approach and a growing ecosystem of plugins.
Linktree - Tons of distribution with creatives and small businesses. Now, they have a marketplace offering where they match their existing user base with new social or commerce services.
Snowflake - What started as an analytics database has also become a deployment target for specific data workloads or applications. Snowflake is also the linchpin of the Modern Data Stack, which companies build complements to.
SpaceX (and soon Starship) - Starship will revolutionize how we take payloads into space, unearthing applications from communications (Astranis) to in-orbit manufacturing (Varda). Although they may have already built the killer application with Starlink.
There’s another dynamic that fascinates me with platforms or would-be platforms. How can they enable forms of new entrepreneurship (not necessarily venture-backed)? You could theoretically build out a small business on the supply-side of Replit Bounties. One could start a new physical retail store 100% on Faire. There’s precedent because this effect has already played out with companies like Airbnb, Uber, Shopify, and others.
As noted, I missed many things so please reach out if other examples come to mind. You can find me at aashay [at] haystack [dot] vc.