I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
Users endorsed model routing for managing AI inference costs and economic value because the thesis has aged well and tackles real engineering problems.
"How do you predict the intelligence of an output with less intelligence than it costs to produce it?"
Model routing is an extremely hard problem.
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
My thoughts on the future of model routing and AI: - We have not even scratched the surface of runaway inference costs - Solving this requires intelligent model routing, especially as the inference landscape continues fragmenting. This is a *hard* problem. - Naive solutions (turn-based routing, session routing) fail; routing successfully involves managing multiple cost surface areas in concert. - Getting routing right means a more diverse market of providers, more power for consumers, reduced ecological impact, and improved effectiveness.
More in the full essay:
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
do you think loops are going to be cheap?
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
"You won’t need to be a power user to spend a shit ton of money on AI. You will just need your AI agent to be a power user to spend a shit ton on AI."
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
"A token is not a unit of value."
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value
@MatthewBerman Enjoyed writing this—and honored to have you on board Matt 🤍
@MatthewBerman I am glad more and more people are starting to tackle the actual real problems rather than building wrappers and “agents”
@MatthewBerman @tomas_hk 👏 good write up @tomas_hk
just keep shipping. everyone needs this, whether they know it or not
@TheAhmadOsman frontier + workhorse + local models are the way to go. it's not one or the other, it's all of the above.
@MatthewBerman Model routing is a dumb idea which will never make sense outside the economics of free llm access.
The obvious solution to the problem, for business users and arguably consumers, is explicit task delegation which carries the advantage of smarter context management.
@MatthewBerman @tomas_hk Model routing is a great start, but just one of the tools that companies will need to employ
Elaborated on this here:
@MatthewBerman @tomas_hk cant believe i was doubting this whole category two years ago
the thesis aged better than my portfolio
@MatthewBerman Brandolini's Law - AI edition.
@MatthewBerman @tomas_hk This is vital to token savings and ensure the best model is used for each prompt/action
@MatthewBerman depends if cheap means low price or low effort
@MatthewBerman @tomas_hk I wonder if routing or just dynamic thinking effort. Are people really going to give up Codex to use Qwen rather than OpenAI models tuning down to low/mini/nano level?
@MatthewBerman @tomas_hk $500m a month on tokens and people still act like it is a zero-sum game
curious how the market splits between routing and native inference over time
@MatthewBerman Consider this problem solved. We’re engineers; solving things is what we do. Now, let's turn that focus toward tackling information overload.
@MatthewBerman @tomas_hk Super interesting read, I had never even considered routing to be the control theory mechanism for intelligence but Tomas laid it our perfectly. Really excited to see how the space develops!
@mokumkiwi @MatthewBerman thanks for checking it out
I made an angel investment in @tomas_hk two years ago, believing model routing would become a massive business.
Now companies are spending $500m per month on tokens. Increased quality and decreased cost are not mutually exclusive.
Here's Tomas' essay about how Model Routing Will Control the Future of Economic Value
https://briefing.forwardfuture.ai/p/model-routing-will-control-the-future-of-economic-value