DeepSeek Shows Us Optionality is Vital - The Weekly #17

DeepSeek is still taking up a lot of column inches around the AI space as more people are coming to realise what the introduction of the R1 model means. Once again, I'm not going to dive deep into its pros and cons but rather highlight the fact that it's yet another model. (I might coin this as YAM moving forward.) In last weeks newsletter I talked about building options into your workflows and not getting boxed into a technical corner and I'm going to develop that idea a little further this week.

It's not just about having the freedom to swap one LLM out for another, but rather having a collection of them at your disposal. You might have picked up the fact that the R1 model is trained to be excellent at reasoning, just like OpenAI's o1.

What does this mean?

These models can use multi-step thinking processes before generating output. They will "think before they speak". This also makes them excellent at solving mathematical problems and coding as they actually work through the problem.

So are they good for everything?

No. If you want to use an LLM to help you craft the opening paragraph of a murder mystery novel, you would be better off using something like Anthropic's Claude 3 model. And here's the point. All models on the market are better at some tasks than others. Certainly, the larger models we are now seeing are more capable in general, but on the whole, you should test models for your particular task and choose the one that gives you the best output.

I'll dig into this concept more in an in-depth article in the next few weeks.

This week's email is brought to you by Kajabi. A leading creator platform that's currently offering a 30-day free trial. I personally use this platform on another project I run and can highly recommend it.

Curated News

AI Platform Palantir Makes Huge Financial Gains

Palantir Technologies, the American data software company, is making significant strides in the AI sector, particularly in national security. The company recently projected 2025 full-year sales of up to $3.76 billion, surpassing Wall Street's expectations by $240 million. This optimistic forecast led to a 24% surge in Palantir's market valuation, elevating it above established giants like McDonald's and Disney.

UK Science Minister Claims DeepSeek Advancements are Good for Britain

Lord Vallance (remember him?), the UK's science minister, highlighted the positive implications of Chinese tech company DeepSeek's advancements in artificial intelligence (AI) for Britain. He emphasized that DeepSeek's chatbots exemplify a shift towards a more competitive AI landscape, moving away from dominance by a few major players. This environment fosters innovation, allowing emerging companies to make significant contributions.

AI Helps Save Patient with Rare Condition

In a remarkable fusion of technology and healthcare, an artificial intelligence tool has been instrumental in identifying an existing medication that saved the life of a patient suffering from idiopathic multicentric Castleman's disease (iMCD). This rare and often fatal condition has limited treatment options, making this development particularly significant. The AI system analyzed a vast array of approximately 4,000 existing drugs to pinpoint a viable treatment, showcasing the potential of AI in accelerating drug repurposing and offering hope for patients with rare diseases

Upcoming AI Events

Big Data World
London ExCel, London, Mar 12-13

AI UK 2025
QE2 Centre, London, Mar-17 & 18

Generative AI Summit 2025
Novotel West London, London, Mar-31 Apr-2

Thanks for reading, and see you next Friday.

Are you looking for a job in the tech sector? You'd do well to check out all the available roles on Wellfound. Wellfound is a platform that connects startups and tech companies with job seekers. It offers over 130,000 remote and local job listings, allowing candidates to apply with a single profile and view salary and equity details upfront.

You can also follow me on social media:

• LinkedIn
• Bluesky
• Instagram