“You Don’t Have the Flexibility to Hot Swap”
On a recent episode of The All-In Podcast, Chamath Palihapitiya and David Friedberg discussed a major challenge facing companies adopting Large Language Models (LLMs): the inability to swap out models as technology advances.
David Friedberg asked where technology companies or investors should start in order to take advantage of rapid LLM innovations.
Chamath explained:
“The first is you have to build a shim, and I think the reason why a shim is really critical is that there’s so much entropy at the model level. What this should show you is you can’t pick any model, and the problem is that the people that manipulate these models—the machine learning engine and whatnot—become too oriented to understanding how to get output of high quality using one thing.
It shouldn’t have been the case that we have engineers that can only use Sonnet—that’s the Anthropic model, right? It shouldn’t be the case that people can only use OpenAI, or people can only use Llama. Right now, that is kind of what we have. You don’t have the flexibility to ‘hot swap’ as models change.
So if you’re starting a company today, the first technical problem I would want to solve for is that. Because tomorrow, if it’s R2 or Alibaba’s model or Llama, I would want to be able to rip it out and put it back in and have everything work. And right now, we can’t do that.”
He’s right—but only if you assume companies don’t already have a solution.
Today, most companies integrate a single LLM directly into their applications, essentially locking themselves into that model. If a better, faster, or cheaper model emerges, switching requires months of development work, leading to rework and technical debt.
But this problem is already solved. Krista provides exactly what Chamath is describing: a shim—a layer that abstracts the complexity of different models so businesses can “hot swap” LLMs instantly.
The Risk of LLM Lock-In
The AI space is evolving too quickly for companies to commit to a single LLM provider. Locking into one model creates major risks:
- Obsolescence – AI models improve rapidly. If you’re tied to one model, you’ll miss out on better performance and new capabilities.
- Cost Increases – AI vendors change pricing structures over time. If you’re locked in, you have no negotiating power.
- Technical Debt – Replacing an embedded LLM later requires rewriting workflows, APIs, and retraining users.
The Solution: Krista’s Instant “Hot Swap”
Krista eliminates these risks by providing true multi-LLM orchestration. Instead of hardwiring a single model, Krista allows companies to dynamically switch between multiple LLMs based on business needs, cost, or performance—without rewriting code.
How Krista Enables Instant LLM “Hot Swaps”
Unlike traditional integrations that require an extensive Software Development Life Cycle (SDLC) to switch models, Krista makes it effortless:
- Pre-Built Multi-LLM Integration – Krista connects to OpenAI, Anthropic, Google Gemini, Meta’s Llama, Alibaba’s model, and open-source LLMs like Mistral—all in one platform.
- No-Code Hot Swap Button – With Krista’s no-code backend, swapping LLMs is as simple as pressing a button. No development work. No SDLC. No reengineering prompts or workflows.
- AI Model Routing – Krista can select different LLMs based on accuracy, performance, or cost—ensuring optimal AI utilization.
The Competitive Advantage of Staying Flexible
By implementing Krista, organizations gain:
- Instant Model Flexibility – Adapt to new AI advancements in real-time.
- Optimized Cost Control – Choose models based on pricing and efficiency.
- Future-Proof AI Strategy – Avoid expensive rework by staying model-agnostic.
AI’s Future is Modular, Not Monolithic
Chamath is correct: the future of AI is not about picking one model—it’s about creating a shim that allows you to switch as innovation happens.
With Krista, you don’t just get a workaround—you get a fully realized “Hot Swap” capability that ensures your AI strategy remains agile and cost-efficient.
Don’t let LLM lock-in slow you down. Explore how Krista gives you the power to swap models instantly and stay ahead of AI innovation.
Sources:
All In Podcast – DeepSeek Panic, US vs China, OpenAI $40B?, and Doge Delivers with Travis Kalanick and David Sacks
