LLM agents and integration dead-ends - by Nathan Lambert

When is GPT4 going to schedule my meetings? What is stopping it?

Thank you for reading and for your support of Interconnects, consider upgrading to paid, if you can, for access to all my work. The practical blocker of deploying ChatGPT for most immediate business writing is the resulting copying of data back and forth into the interface. Bringing this barrier down to enable direct integration of generative AI is the biggest immediate roadblock in generative AI unlocking huge economic value (yes, the >$ 4 trillion per year of profits that McKinsey forecasted across only 63 use cases). Curiously, it seems like there is currently an appetite for these…

related reading

• Building Effective AI Agents \ Anthropicanthropic.com
• Building Effective AI Agents \ Anthropicanthropic.com
• LLM Powered Autonomous Agents | Lil'Loglilianweng.github.io
• Guardian Angels: LLM Personalization for Productivity and Security · Gwern.netgwern.net
• LLMs Are Not All You Need | Pineconepinecone.io
• What We’ve Learned From A Year of Building with LLMs – Applied LLMsapplied-llms.org
• The Problem With LangChain | Max Woolf's Blogminimaxir.com
• What CEOs Need To Know About The Next Frontier Of AI: AI Agentsforbes.com
• Things we learned about LLMs in 2024simonwillison.net
• The lethal trifecta for AI agents: private data, untrusted content, and external communicationsimonwillison.net
• How fast is AI improving? - AI Digesttheaidigest.org
• Why I'm Betting Against AI Agents in 2025 (Despite Building Them)utkarshkanwat.com