PrecisionThe Storm Behind the Models
Hello and welcome to the latest Precision Weave newsletter.
In this edition we are focussing on the ways that ChatGPT and Claude differ in the tasks they do best and what it means for you. But first, I wanted to remind you I will be giving a briefing on Practical AI For Small Business at venues across the West Highlands starting next month. Tickets are available now on Eventbrite.
- Fort William – Tuesday 28th April – 11:00am at UHI Fort William Campus – book here
- Skye – Tuesday 5th May – 2:00pm at UHI North, West and Hebrides, Portree Campus – book here
- Oban – Tuesday 12th May – 2:00pm at SAMS Dunbeg, Oban – book here
I will also be attending the West Highland Vision conference which is being held in Oban on Thursday the 12th of March. You can book here. Please do come and say hello if you are there.
All the best,
Adrian
New on the Website
Claude vs ChatGPT: What Actually Matters for Your Business
The Punctuation That Makes AI Listen
Editorial
At long last, winter is slowly passing into spring here in Fort William, but there has been a storm raging in the world of AI!
You may have seen that Anthropic, the company behind Claude, objected to their models being used in certain military situations. The White House responded by designating Anthropic a “supply-chain threat” to national security. OpenAI were happy to accept the Pentagon’s terms, but the public backlash was severe. Claude surged in popularity and OpenAI were subsequently forced to “clarify” their contract with the Government.
None of this has any direct effect on how you and I use AI day to day. But it seemed like a good moment to lift the lid on how Claude and ChatGPT actually differ. The models have evolved so fast that any of the big names can handle tasks that would have been cutting edge this time last year. The real differences now aren’t about which model is “better.” They’re about how each one is trained and the kinds of tasks they’re optimised to perform.
I personally use Claude about 98% of the time, mainly because it’s great at writing code. But when I dug into it for this issue’s blog post, I discovered the differences run much deeper than I’d realised and one is optimised for what people prefer in the moment while the other is optimised for consistency. Check out the post for the full rundown.
I’ve also added a post about improving your prompts using markdown formatting. If you’ve ever wondered what those hash symbols, stars, and backticks you see in some prompts actually mean, this one’s for you.
As ever, please do reach out if you’ve got any questions.
Happy prompting, and enjoy the spring sunshine.