Is Your AI Assistant Playing Hard to Get?
Ever had your AI sidekick ghost you mid-conversation?
It's not you, it's them - or more precisely, their overloaded servers.
Ah, Availability…next to the dreaded I of the CIA triad…
In the world of AI, all roads might lead to Rome, but some roads are more congested than others.
When your go-to AI service hits a traffic jam, it's time to take a detour.
Here's how to keep the conversation flowing:
Diversify your AI portfolio: Set up accounts and API keys with multiple LLM providers (big-name providers, plus OpenRouter for Claude Sonnet 3.5 access without quotas.
Configure for quick switching: Organize your API keys and endpoints for easy swapping in your code or tools (but keep ‘em out of GitHub!)
Know your options: Familiarize yourself with the quirks and strengths of different models. Claude might excel where GPT4o stumbles, or vice versa.
Monitor service status: Follow your preferred AI providers on social media or set up alerts for their status pages. OpenRouter has regular database issues…but they are generally quick to recover.
Practice the switch: Regularly alternate between services to familiarize yourself with each interface and API. This is not as dumb as it may sound; your favourite standby model may get renamed or retired after just six months of service!
Remember, in the AI world, loyalty doesn't pay. Being promiscuous with your LLM providers ensures you're never left hanging when you want to travel fast.
What's your favorite backup AI service when your primary choice decides to take an impromptu siesta? Any hidden gems you've discovered in the LLM landscape?
Cheers, Craig