Eye On A.I. · Oct 16, 2026 · 55m · 47m saved
With Iskander Malikov, Co-Founder and CEO of Trioma
Trioma connects 65 AI models behind one interface and decides for the user which model, or which combination, should handle each request.
Iskander Malikov, Co-Founder and CEO of Trioma, told Eye On A.I. that the price per token most buyers compare is the wrong number. Most people pick one chatbot and pay a flat monthly fee. Iskander argues the real cost of AI is the total spent across however many attempts it takes to get a usable answer.
"You're looking at how many iterations it took me to get to a result that I'm satisfied with... you realize that the price of token is meaningless if the end result is like 10 times cheaper."
Iskander's UAE-based company had its official launch the day before this conversation, after building since January and opening the product to early users in August.
The full interview is covered here so you can skip it. 55 minutes of audio, 8 minutes of reading.
Here are the 7 arguments that matter.
This Scuttlebutt Summary is available in full to premium subscribers. If you're not one already, please consider subscribing to keep reading.
Key Takeaways
Most users want the finished result, not the AI's reasoning, which is why Trioma picks models for them
Token price is the wrong metric: the real cost is the total spent to reach a usable result
Routing a task to a different model gets around another model's refusals
Trioma lets users buy into its GPU buildout and earn credits that offset their AI bills
Flat AI subscriptions do not fit AI's costs, because every request burns real compute
That gap is why frontier labs keep adding usage limits
Agentic coding threatens software like Figma and Photoshop
A co-founder built a working messenger app in about 10 minutes
1. Results, Not Reasoning
Craig Smith, who hosts Eye On A.I., asked Iskander what Trioma actually is. Iskander started from what users want, rather than which model they use.
Iskander Malikov built Trioma on the bet that users care about the output, not how the AI got there
"What we've realized is that most people actually want certain thing to get done. So, when they interact with AI, they are mostly interested in the result itself, not so much the process, the reasoning, how the AI is processing your specific prompt." — Iskander Malikov
Iskander described this as a shift from AI as search, where a user asks a question and reads an answer, to AI that delivers a finished piece of work. Trioma is built in three layers to serve that.
"Trioma is a vertically integrated AI ecosystem where there are three separate layers... Then there's the mid layer, where we have 65 different AI models, as well as the orchestration layer that connects them... and then there's the bottom layer, which is GPU, chips that power that compute." — Iskander Malikov
The top layer is the interface, on the web and desktop, with iOS and Android apps coming. New models are added within hours to days of their release. The expensive part is tuning the orchestration engine so a new model actually gets used for the tasks it does well.
