I remember reading about the “hybrid start ups” from HBR; I’ll be linking to it in a future blog post. I witness an interesting development; the biggest-tech companies are making “risky investments” (or bets in VC parlance) in ideas outside their core-domain, but with overlapping interest. OpenAI was funded by Microsoft, which is integrating it into “co-pilot”. It is not perfect, and I wouldn’t base my opinions on a demo.
Nevertheless, it is a win-win situation for everyone involved. Companies avoid the intense scrutiny that comes with the AI missteps, and the “start-ups” get the resources they need.
Here’s the blog post that is a bellwether for this:
Why? Because large companies can’t innovate the way startups can.
- Slow product cycles from layers of bureaucracy
- Extreme reputation risk (when things inevitably go wrong)
- Disruption to existing business models (e.g., Search)
- Distribution to a small but important new audience (e.g., LLM builders)
- Incentives to make risky bets (e.g., startup equity)
The hardware is an issue:
From the blog post:
Startups fight to request quota on the cloud providers (and others, building their own data centers). At the lowest level, many of these models are tightly coupled to the underlying hardware, and I believe we’ll only see more application-specific chips (read: transformer-optimized architectures). Cloud hyper-scalers have started building this ability over the last few years, and Google and Apple have done so via chips for mobile. Even Meta has built out this capability with its partnerships in VR headsets. For what it’s worth, Meta has also worked to abstract the underlying GPUs (re: commoditize) out of PyTorch.
We will see further development of an “application-specific” chipsets; who knows, perhaps dedicated GPU’s for mobile phones? AI will increasingly be refactored in mobile operating systems (like iOS via Stable Diffusion) and Android through Google Pixel line of phones incorporating Google designed and Samsung manufactured chipsets (Exynos). These are interesting times. AI won’t be “democratised”, as you might hear the mainstream tech publications crow. It will be a test bed to create control structures for their other businesses.
RadOncNotes 