Welcome To The Era Of AI Assistants
Something has changed. There’s a growing optimism in the air. Working in the field of conversational AI has never been more exciting. I’m thoroughly convinced we live in the era of AI assistants.
The interest from outside is growing, too. I started this newsletter end of last year as a way to share my thought and ideas. Some practical, some more philosophical. The amount of attention it has generated over the past months has exceeded all my expectation. So, to all my early readers, I would like to express my gratitude.
And to all my new readers, I would like to say welcome.
A small recap of the last months
There were quite a few product launches and announcements. Here’s a small rundown:
OpenAI launched GPT-4
Microsoft launched the New Bing
Google launched Bard
Anthropic launched Claude
Character.ai $150 million Series A funding round led by a16z (link)
Anthropic's 4-year plan to take on OpenAI (link)
Elon Musk says he will launch rival to Microsoft-backed ChatGPT (link)
Mark Zuckerberg wants to ‘introduce AI agents to billions of people’ (link)
It is also the year of open source language models, a few in no particular order:
Some of these were trained by fine-tuning LLaMA, Meta’s AI language raw model that was leaked on 4chan only 4 weeks after it was announced by researchers. The cost of re-training can be done for no less than $300.
Chatbots on the loose
Unsurprisingly, it is also the first year we’re seeing chatbots run amok.
We had Bing’s alter ego Sidney saying some pretty weird stuff (link)
Replika’s AI companions sexually harassing people, and leaving others heartbroken (link)
The first death associated with a LLM-powered AI assistant (link)
And a bunch of other stories that suggest this technology is evolving faster than society’s ability to adapt.
In the race to be first-to-market some are willing to cut corners. But the old Silicon Valley adagium ‘move fast and break things’ hasn’t aged well. Developing experimental conversational AI solutions asks for a more thoughtful, and frankly more ethical approach: there is a deeply psychological dimension at play, introducing a unique set of risks that need to be mitigated.
Many have called for central oversight — and even a worldwide pause on developing general-purpose AI has been suggested.
What to expect from this newsletter
My goal with this newsletter has always been (and will continue to be) to write about interesting stuff.
In an age where the cost of producing low-quality content is going to zero, the need for high-quality content will only grow. I’m convinced of that. People long for good stories. In-depth articles. Real thoughts.
To all my new readers, here are some of my latest publications:
And many will follow. For now, thank you for your continued support and have a great day!