You’re also cutting off developers who care about the cybersecurity of their agents and don’t want to point them to random websites that could contain dangerous prompt injections, as well as people who want to understand where they’re directing the agent and why before doing so
I think the biggest thing is to not give it access to anything like a shell (obviously), limit the call length, and give it a hangup command.
Then you tell it to just not answer off the wall questions etc. and if you are using a good model it will resist casual attempts.
I don't see being able to ask nonsense questions as being a big deal for an average small business. But you could put a guardrail model in front to make it a lot harder if it was worth it.
in general these types of attacks are still difficult to solve, because there are a lot of different ways they can be formulated. llm based security is still and unknown, but mostly i have seen people using intermediary steps to parse question intent and return canned responses if the question seems outside the intended modality.
That's not what I meant. I wasn't talking about replicating the popularity (that never materialized), but about getting on the hype train of selling people nothing for something.
It could be about cost: big LLMs don’t work locally like most games, and even with server hosting, the context for a well-detailed and non-textual game would grow very fast, so processing it would become slow and/or expensive. These are the current limits for truly generative games.
As for generated content in regular games, I don’t see an issue if the content is high quality and free of errors. People don’t like low-quality content regardless of who generated it - human or LLM. It’s just that there’s currently more bad content coming from LLMs, that’s all
Generating text & audio is relatively cheap. If users paid a $5/month subscription you could generate a couple hours of dynamic npc audio per month and still make a tidy profit.
I have to imagine someone is looking into that, sandbox style games with hundreds of characters who have unique personalities and respond to any input and remember all your interactions... that would be amazing.
I’m trying to use the CLI whenever possible - it’s much easier to install and can be used by both me and the agent. For example, gh seems much easier than installing and setting up an MCP server connection, and it’s more human-readable in terms of what the agent is calling and what it’s getting in return.
For other integrations, I first try to find an official or unofficial CLI tool (a wrapper around the API), and only then do I consider using MCP
reply