How LLMs for chat turned AI into a general-purpose information-processing tool, as AI undergoes a transformation much like how chips gave rise to CPUs and GPUs
This coverage identifies conversational LLMs as an interface shift: a single model can be applied across many language and information tasks rather than being built for one narrow use. Follow-on coverage focused on making those systems more controllable and interactive, a practical requirement for use beyond open-ended chat.
The arc also broadened from cloud chat to smaller models: reporting on Phi-3 running locally on phones and PCs tied the same general-purpose capability to responsiveness and privacy constraints.
First-order effects
Chat becomes a common entry point for applying AI to varied information tasks, lowering the need for users to learn separate task-specific AI tools.
Model developers and product teams must turn broad language capability into controllable interactions before it can be reliably embedded in products.
Second-order effects
Competition shifts toward deployment choices as well as model quality: local small models can serve uses where responsiveness or privacy matter, while larger systems remain suited to other workloads.
Software providers have an incentive to place AI inside existing work surfaces rather than leave it as a standalone conversational destination.
Third-order effects
If general-purpose models continue to improve, differentiation is likely to move from a model's raw breadth toward workflow integration, control, and the compute architecture behind each use case.
The CPU/GPU analogy points to a more specialized AI stack over time: broadly useful models may coexist with purpose-built models and heterogeneous hardware rather than replace them outright.
The trend: Conversational LLMs are becoming a general interface layer for software, while deployment and workflow constraints drive specialization underneath.
Piercing the AI hype, Paul with an interesting take that says: “We are already at the tail end of the current wave of AI. We are bumping against many of its limits...”
Some papers mentioned: • Population Aging and Economic Growth https://www.nber.org/... • Automation and New Tasks: How Technology Displaces and Reinstates Labor https://www.aeaweb.org/... • Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks https://arxiv.org/...
As a postscript, while we say and believe this, we are also actively investing in our venture partnership around this thesis. We can be accused of talking our book, but we think we are mission-driven with thoughtful ideas about technology and the potential for human flourishing.
The piece by @defrag and I for our SK Ventures is up. It's called “AI isn't Good Enough”. It is on structural imbalances in the US workforce, demography, the risks of so-so automation, and the potential for human flourishing. You can read it here https://skventures.substack.com/ …
We go into some current developments in AI, like the rise of RAG (retrieval augmented generation), which is timely in the context of OpenAI's latest moves. And, for econogeeks, dive into some automation models from Acemoglu and Restrepo, before making a Led Zeppelin Zoso claim.
We argue in the piece that AI needs to be much better or, weirdly, much worse, to avoid the so-so automation trap, where we have massive economic dislocation without compensating productivity gains. We also map out our views of what the next decade could look like.