After one year of implementing AI features for various businesses, I share my perspective on the mistakes I see companies making with LLMs, but also which strategy to adopt.
2023 has been the year of massive achievement in artificial intelligence, especially in Natural Language Processing with Large Language Models (LLMs).
With the apparition of Generative AI and the impressive performance that came with it, most companies revised their strategy to get AI in their product.
In addition, start-ups with the term “AI” embedded in their name have been emerging in all domains with one main goal: finding a problem that GenAI could solve.
We entered an era of AI hype: which company will produce the best open-source LLM — have the best product demo video on Twitter — and replace all its strategies to incorporate AI in it.
We are right now in a bubble. When this bubble will pop, most of the companies using LLMs will disappear…
But it doesn’t have to be the case.
With the correct strategies, the right expectations, and a comprehension of what it takes to develop AI features, I’m convinced it is possible to provide a ton of value for customers and thrive as a business.
As a freelancer, I have been advising companies and developing features using LLMs in various domains: HR, sales, short-term rents, AI start-up, …
In this article, I share my raw experience as an NLP engineer during this hype and my opinion on mistakes I see companies making.
I then explain the reality behind an LLM, how it works, and what it takes to develop AI features.
I finally propose my vision of the right strategy to adopt to make the transition to AI a success.
The story of an NLP engineer through the AI hype
“GPT-3.5’s here. We need to surf the wave!”
I started being involved in LLMs back in January 2023.