GPT Function Calls come with native support for planning tasks in the presence of multiple available tasks, tools, and actions. This is an emergent behavior and greatly simplifies building LLM agents, over existing frameworks such as Langchain.

GPT Functions Are Native Task Planners

We do an honest review of the new function calls feature from OpenAI and analyze it from the perspective of building LLM agents. We explore the benefits of function calls, and how they can be used. Lastly, we mention some of the current limitations of function calls. 

Building Agents with GPT Function Calls - Langchain Killer?

We explore why most ChatGPT Plugins and LLM agent projects feel underwhelming, and uncover some challenges developers face in building LLM agents. The challenges are mainly in four categories: 1) latency, 2) reliability, 3) UX, and 4) trust / security. We also argue that many LLM agent projects are doomed to fail because natural language is an inferior user interface to GUIs in many situations.

Challenges of Building LLM Agents and Plugins

In this guide, I will go over key concepts that any ChatGPT developer should know:
- what is a ChatGPT plugin (in plain English) - how ChatGPT plugins are different from iOS apps - how ChatGPT plugin dispatch model works - gotchas and issues with building and using ChatGPT plugins - whether you should build a ChatGPT plugin

ChatGPT Plugins - A Step by Step Tutorial & Best Practices - (Part 1)

We cover 17 best practices for optimizing cost with Pinecone, specifically for the newcomers to vector databases as target. These practices will save you potentially tens of thousands of dollars. The advice is grouped into four buckets: 1) general tips, 2) application-level best practices, 3) infrastructure-level best practices, as well as 4) advice specific to the paid tier.

Pinecone - Cost Optimization & Performance Best Practices

Stanford University's release of the Alpaca model - which basically showed that an inferior foundational LLM can mimic a superior model cheaply - has massive business and science implications. Though "self-instruct" - i.e. the pattern of fine-tuning models with auto-generated demonstrations/training data - is nothing new, Stanford showed that it could be done at the scale of LLMs, and that it only costs $600! If Stanford's results are valid (and that's a big "if"), here are the major implications.

The Profound Implications of Alpaca on OpenAI and Competition

Is AI adoption actually beneficial for tech stocks? While enthusiasm for AI has piqued investor interest in tech stocks again, i the long-term impact of AI on tech valuations may not be as positive as initially thought. In this article, we'll explore why AI adoption may actually be bearish for the average tech stock, particularly for IaaS, PaaS, and SaaS companies. We'll delve into the risks associated with rapid AI advance, such as execution risk, reduced cash flow visibility, and a shift in data and compute gravity. Additionally, we'll discuss how the rise of AI may reduce the demand for white-collar workers, leading to fewer seats and less revenue for SaaS companies. Ultimately, while the impact of AI on tech valuations may not be immediate, strategic mishaps will inevitably lead to re-pricings over multiple earnings cycles.

Here's How AI Can be Bearish for Tech Stocks

Are you interested in AI but feeling overwhelmed by the technical jargon? Learn the foundational ideas behind ChatGPT, GPT3, and Large Language Models (LLMs) in plain English. Discover how these models generate text and the "secret sauce" that makes them so powerful, and how InstructGPT model differs from GPT3.

John Hwang@nextworddev

Recent Articles