Last week I dug into a new emerging challenge for AI investors. As existing portfolio company ambitions grow alongside emerging AI, they'll have to find a way to both find the next post-ChatGPT Databricks and Hugging Face without competing with their existing investments.
@mattlynley
-
PicPlz Instagram Drama Echoes in Early iPhone App Wars
By
–
If you remember what happened with the PicPlz/Instagram weird drama back in the early days of the iPhone, this one should sound a little familiar. Full story can be found here:
-
AI Gold Rush: Navigating Investment Conflicts in LLM Startups
By
–
Today's issue digs into a growing dilemma for investors: investing in the new gold rush in AI, while avoiding startups competing with their portfolio companies that are breaking into LLMs from the LAST AI gold rush
-
The iPhone moment required ecosystem maturity before mainstream adoption
By
–
The so-called "iPhone moment," though, was actually a lot more drawn out than people realize—we needed a foursquare, an instagram, an uber, and a few others for it all to make sense. And there was a tsunami of apps in all of those categories, most of which are lost to time.
-
MLops Stack Race: Investment Dynamics and Market Competition
By
–
It's not so dissimilar from the last mad dash to claim the MLops stack. And investors looking to tap into that wave have to still follow the general house rule of venture: don't invest in competing companies. And we actually saw this same thing happen during the "iPhone moment."
-
Startups Rush to Build LLM and Diffusion Model Deployment Stack
By
–
Those same companies have established themselves in the billion-plus club and now eye enabling deployment of LLMs/Diffusion models. But there's a whole NEW gold rush by startups looking to pick off every piece of the language/diffusion model deployment stack.
-
New AI Wave: Investors Navigate Competition Beyond MLops Stack
By
–
Now that we're entering a NEW AI hype wave, investors are finding themselves with a growing dilemma: avoiding competition with the last one. In the last wave (we'll call it 2019-2022) a whole ecosystem of startups sprouted around picking off individual pieces of the MLops stack.
-
Developers competing on AI performance metrics shift toward softer approaches
By
–
For months some developers have been trying to one-up each other using performance evaluation metrics. But some companies may look to use a softer touch. (Free issue)
-
LLM Leaderboards Fail as Product Quality Indicators
By
–
While the open LLM leaderboard provided (and still provides) a good starting point for a lot of companies exploring integrating LLMs into their products, the one-upmanship of it seems to be less and less of an indicator of whether a model is "good" for a specific product.
-
Iterative Parameter Tuning in Machine Learning Model Optimization
By
–
In particular, they "evaluate" by twisting the knobs, so to speak, until they land on a good user experience. If you've worked in data science or ML, that should sound like a very familiar experience.