Many users endorsed the claim that enterprises will lock into single AI model stacks because of real challenges with agent harnesses, evals, and workflow capture, while a few replied with sarcastic dismissals.
Based on 14 visible X reactions from 47 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Matches what we see in enterprise deployments, one addition: the moat isn't the evals — it's that nobody captures what the "workflow" was supposed to do anywhere outside the model stack. Data ports, facts port. Intent doesn't — it lives buried in prompts and harness code. That's why re-verification is a six-month project instead of a canary deploy. The 75% lock-in feels right as the default, but it's a choice, not physics. Capture agentic workflow intent above the harness, and verify in production economics — did it complete, what did it cost, what was it worth — and a model swap becomes something you watch, not re-test. The orchestration layer needs design genius. The verification layer just needs the discipline to measure outcomes — that's what we're building at Agently AI an agentic workflow intelligence.
@nikesharora Absolutely spot on from our observation too while iterating through agent harness, it gets tuned to a specific model despite every effort to generalize it. Newest, more capable model on an existing harness does not mean more capable harness.
@nikesharora @nikesharora Thank you for putting this so eloquently, This mirrors exactly what we’ve learned building Tuskira’s agent stack, The harness argument is the one people underestimate. (1 of many)
@nikesharora bro built a whole thesis and the conclusion is 'switching models is like moving apartments but your furniture is made of vibes' 💀
Venture capitalist Brad Gerstner publicly endorsed Arora's assessment.
@nikesharora They might, but they’ll give up when they realize it doesn’t make the beer taste better
@nikesharora 🎯🎯🎯
Summary: I spent time trying to figure out this orchestration layer problem, can we design a multi model architecture in the long term. The more I dug in the more I understand that trying to build an abstracted layer is hard. As agentic activities increase and agent chaining and complex tasks get assigned to AI it will become harder to move between models. There is a reasonable probability that 75% of the enterprises will build their implementation of the solution to their core problem around one model "stack". Token price reduction by 90% is the solve and mobility between models from the same frontier lab! Evals, harnesses, cache memory are the moats and I don't see models providing simple abstraction to those. I know there are efforts to do this out there, the long term solve for orchestration if it works will need to be "Claude code" level of design genius. Here's a chat with Fable @HamzaFodderwala had. **Why abstraction looks easy.** Models are stateless — every API call is weights + a prompt assembled at runtime. Everything the model "knows" about you — memory, documents, history, tools — is injected into the context window by software outside the model. So in principle, all your state already lives outside the weights. The catch is what "state" includes. **Layer 1 — Data (fully portable).** Enterprise documents, tickets, logs. Retrieved via RAG: text is chunked, embedded, stored in a vector database (Pinecone, pgvector), and relevant pieces are fetched into the prompt per query. The embedding model is separate from the LLM, so this layer is genuinely model-agnostic. Already solved. **Layer 2 — Memory (portable in principle).** Systems like Mem0 and Zep sit between the app and the model: after each interaction they extract salient facts ("user prefers X"), store them as plain text, and inject the relevant ones into future prompts. Because the artifact is natural language, it reads into any model. Facts port. **Layer 3 — Orchestration/routing (works, but only for shallow tasks).** Gateways like OpenRouter and LiteLLM normalize API differences and route each request to the cheapest capable model. This is the fungibility layer being furiously built. It genuinely works for one-shot, verifiable tasks — classification, extraction, summarization — which conveniently are the tasks where cheap models suffice anyway. **Where it breaks — the non-portable state.** Four things stay behind when you switch: - **The harness.** Prompts, tool schemas, and guardrails are tuned to one model's quirks. An agent must get every step right, so reliability compounds: a model that's 98% reliable per step completes a 50-step task about a third of the time; at 90% per step, it almost never finishes. Swapping models costs you a few points per step — the difference between an agent that works and one that doesn't. - **The evals.** Swapping means re-testing everything and re-fixing every regression. The real switching cost isn't data migration — it's re-verification. Nobody has abstracted that. - **Procedural memory.** Facts port; skills don't. Cached successful workflows and learned workarounds are conditional on the model that produced them. - **Cache pricing.** Provider-specific, worth 75–90% of input costs on agentic workloads. Quiet lock-in. **The labs' angle.** They offer hosted memory, hosted file stores, caching, fine-tuning — every one pulls state from your side onto theirs. The labs will crack memory first, but as lock-in, not portability. Nobody standardizes their own exit door. MCP is the partial exception: it standardizes tool and data access across models, but doesn't touch harness tuning or evals. **Where 3P vendors fit.** Routers are thin-margin commodity plumbing; vector DBs and memory infra are real but small. The two structurally interesting positions: **eval platforms** (LangSmith, Braintrust) — since switching cost equals re-verification cost, whoever industrializes cross-model testing actually enables fungibility.
Many users endorsed the claim that enterprises will lock into single AI model stacks because of real challenges with agent harnesses, evals, and workflow capture, while a few replied with sarcastic dismissals.
Based on 14 visible X reactions from 47 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@nikesharora 🎯🎯🎯