OpenAI Dev Day
Extinction event for the Cambrian explosion of AI startups or par for the course at early stage?
OpenAI held its first ever Dev Day yesterday and the startup world is a churning sea of red. 🚨 People have variously described it as a 'Red Wedding', 'Doomsday', or 'Extinction Event' for AI startups. Why? Well, TLDR is that their core products got better and cheaper + they're entering adjacent markets.
OpenAI has built its castle atop a hill and is now getting started on building the moat. 🏰 Most announcements are geared towards making it easier to build on top of their platform and locking in customers. 🔓
👉 GPT 4 now has a Turbo version with a 4x larger context window of 128K tokens, updated knowledge cut-off as of Apr-23, and lower $$ cost. This issues a challenge to other foundation model companies ( 👀 on Anthropic, Cohere, Bard) to improve their cost-benefit equation for customers because GPT-4 Turbo just got leaner, meaner, cheaper. 🔥
👉 Audio API converts text-to-speech and competes directly with the likes of ElevenLabs at 1/5th the price 📢
👉 Assistant API lets developers build 'agent like experiences' within their app and build digital assistants with specific instructions that leverage outside knowledge and can call OpenAI models and tools to perform tasks. Instead of finding a third-party to chain together these apps, you can just use OpenAI APIs for it now. 🤖
👉 Developers can now directly access specialized data without needing to change it for AI models, challenging standalone vector databases 🔢
👉 Users (even those with limited coding knowledge) can build their own custom GPT models for specific purposes and publish them on OpenAI's newly announced GPT Store. This is the first step in OpenAI becoming a true platform and potentially unlocking a developer ecosystem as vibrant as Apple's App Store 🏪
There are several such announcements which underscore that building in AI right now is like constructing a city on the back of slumbering giant - you never know when it's going to wake up and shake you off. That is why, more than anything else, a founder in this space requires tenacity, resilience, and adaptability.
Do I think the set up just got more challenging for many startups? Sure. Do I believe that they're dead in the water, like many hyperbolic tweet-storms would have you believe? Not in the least. The world of venture is built around the belief that David can beat Goliath. OpenAI is undoubtedly a force to reckon with but when you try to be everything to everyone, you leave space for those that focus entirely on winning the market they set their eyes on. This is when really understanding the customer and reimagining workflows will win the long game. Here's to always betting on the Davids. 🥂
The statement "you leave space for those that focus entirely on winning the market they set their eyes on". really resonated with me in a positive manner. I strongly believe that smaller businesses can indeed gain an edge by reevaluating their processes utilizing AI and prioritizing the needs of their clients.
Given the paced nature of the AI industry I'm interested in your insights on how these agile startups can adjust and thrive. I would love to hear more, about your expertise regarding the strategies that may emerge.