Pushing the boundaries...

The famous bridge in the garden at Giverny was one of Monet’s most famous painting. As an artist, he pushed the boundaries and today, we have another Claude that is doing the same thing. Let’s dive in…

The world of Large Language Models (LLM’s) has taken another bold leap forward this week as Anthropic released it’s Claude 3 model in three flavours: Opus, Sonnet and Haiku. In performance benchmark tests that are common in the industry Anthropic have raised the bar.

In laymans language, these results set levels of performance across a wide range of activities that we have not seen to date. Tellingly, it is also the first time that OpenAI have not led these benchmarks. They have been working on releasing ChatGPT-5 which, from people who have seen it, will undoubtedly push the dial forward still further. But to what end? What does the march of the models actually mean for you and me?

Firstly, greater intelligence means greater understanding and less likely to hallucinate. Greater understanding of the nuances of human language means you are much more likely to get outcomes that align with your intentions. Getting rid of hallucinations is, we have been told, inevitable as these models progress and this will feel like a huge leap forward.

The next advantage of more intelligent models is the ability to process massive amounts of data at incredible speed. The genuinely useful AI assistant is becoming more and more of a reality. By that I don’t mean “Alexa, turn off the lights”. I mean an assistant that can still do that, but can also do complex, high value tasks. For example, scouring the internet for the best flight deals to your preferred holiday destination, reading through thousands of reviews of various hotels and attractions to get the best places for you to stay and visit, communicate with these places to get the best deals, and book everything using your preferred payment platform that also takes advantage of any points and rewards that you are entitled to. This is not science fiction, we are on a runway to this within a very short period of time. And it will be reliable, useful and you’ll wonder very quickly how you ever managed without it.

In the world of work, we will start to see AI really influence workflows making us all supremely efficient, doing much more, much more quickly. Such advances will bring huge economic benefits across the entire world, and whilst this will undoubtedly lead to significant job losses, there will be lots of opportunities for reskilling and income redistribution (which is another newsletter topic all on its own!) to ease the pain and hasten what will be a societal change likely to be even greateer in scope (because of the speed at which it will happen) than the Industrial Revolution.

So, here’s to Claude, both the artist and the LLM, who pushed the boundaries and made life more enjoyable.