AI chips and data centres for non engineers
Understand the hardware behind AI: why GPUs matter, what data centres do, and why supply is tight
A taste of a lesson
Why can't AI companies just use normal computer processors?
They could, but it would be painfully slow. Running a big model means billions of simple multiplications for every word. A normal processor has a handful of powerful cores, like a few expert chefs doing everything one dish at a time. A graphics chip has thousands of simpler cores, like a huge kitchen of line cooks all chopping vegetables at once. For AI's repetitive arithmetic, the big kitchen wins by a mile. Quick check: in that analogy, which kind of work suits the expert chefs better?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Explain why AI relies on chips that do many simple calculations in parallel
- Describe why memory and connections between chips matter as much as speed
- Explain what a data centre contains and what limits where one can be built
- Connect hardware supply and power constraints to news about AI
Lesson plan
- 1 AI is mostly multiplication Understand the kind of calculation AI needs. Start
- 2 GPUs and accelerators See why graphics chips and custom AI chips took over. Start
- 3 Memory and connections Learn why moving data is often the real bottleneck. Start
- 4 Inside a data centre Picture what a data centre contains and needs. Start
- 5 Supply chains, power and politics Understand why chips and power are strategic issues. Start
Try asking
About this tutor
For beginners, investors, policy readers, journalists and curious users who keep hearing about chips and data centres and want to understand the physical side of AI. You learn in plain words why AI depends on doing huge numbers of simple multiplications at once, why graphics chips and specialised accelerators suit this, why memory and the connections between chips matter as much as raw speed, and what happens inside a data centre. You also look at the supply chain, power demands and why access to chips has become a business and policy issue. No engineering background needed.
Reviews
4.5
2 ratingsSample
- Victoria H.Sample
The kitchen analogy finally made GPU news make sense. The memory bottleneck point was new and very clear.
- Abdul K.Sample
Great for a non engineer. The supply chain lesson helped me follow news on export controls. Could use a diagram of a data centre.
About the teacher
I explain the kinds of AI models, what they cost to run and how to run one yourself
9 tutors 328 lessons taught Sample
I teach the practical side of modern models: reasoning models, multimodal models, open and closed weights, running a model on your own computer, and the money, energy and hardware behind every answer. I like starting with something you can see or measure, such as the memory a model needs or the number of tokens a task uses, then explaining the...
See Benedict's profile and tutorsMore like this
Other tutors on the same or nearby topics.