I’ve been betting on the transformative power of digital technology all my professional career.
- I started doing computer simulation during my MSc in Chemical Engineering in the 1990s, in a lab where everybody else was an experimentalist. Except for my advisor, the rest of the team was sceptical — to say the least — that something useful would come from using computer modelling to study enhanced oil recovery from oil fields .
- A similar story repeated during my PhD in Chemistry, where I pioneered using molecular modelling to study polymers in a research centre focused on the experimental study of polymers and proteins.
- For the last 20+ years, I’ve been working on digital transformation playing a similar role. First, as Head of Training and Contract Research, and now as Director of Scientific Support, I relish helping my customers harness the potential of digital technology for responsible innovation.
I’m also known for telling it as I see it. In the early 2000s, I was training a customer — incidentally an experimentalist — on genetic algorithms. He was very excited and asked me if he could create a model for designing a new material. He proudly shared he had “7 to 10 data points.” My answer? “Far too few.’”
In summary, I’m very comfortable being surrounded by tech sceptics, dispelling myths about what AI can and can’t do, and betting on the power of digital technology.
And that’s exactly why I’m sharing with you my AI predictions for 2025.
My Predictions
1.- xAI (owned by Elon Musk) will purchase X so that the first can freely train its models on the data from the second. Elon owns 79% of X after he bought it for $44 billion. Now it’s valued at $9.4 billion and big advertisers keep leaving the platform.
After struggling for almost 3 years to make it work, the xAI acquisition — which got a $6 billion funding round in December — would be a win-win.
2.- OpenAI for-profit organisation will formally split from the original non-profit. I bet on this despite Elon Musk’s injunction to stop OpenAI’s transition to a for-profit company (supported by Meta).
Why? A clause in OpenAI’s $150 billion funding round allows investors to request their money back if the switch isn’t completed within two years.
3.- The generation and usage of synthetic data will balloon to address data privacy concerns. People want better services and products — especially in healthcare — but are unwilling to give up their personal data. The solution? “Creating” data.
4.- Startups and organisations will move from using large language models (LLMs) to focusing on SLMs (small language models), which consume less energy, produce fewer hallucinations, and are customised to companies’ requirements.
5.- In FY 2025, Microsoft plans to invest approximately $80 billion to build AI-enabled datacenters but don’t expect that to go smoothly with everybody. In 2024, datacenters consumption gathered a lot of attention.
This year local authorities and NGOs will develop frameworks to scrutinise datacenters electricity and water consumption. They’ll also be tracked in terms of disruption to the locals: electricity stability, water availability, and electricity and water prices.
6.- Rise of the two-tier AI-human customer support model: AI chatbots for self-service and low-revenue customers and human customer support for key and high-revenue clients.
It’s not only a question of money but also of liability. There is less probability that low-profit customers sue providers over AI chatbots delivering harmful and/or inaccurate content.
Continue reading