The Times Australia
The Times World News

.

How a small Chinese AI company is shaking up US tech heavyweights

  • Written by Tongliang Liu, Associate Professor of Machine Learning and Director of the Sydney AI Centre, University of Sydney



Chinese artificial intelligence (AI) company DeepSeek has sent shockwaves through the tech community[1], with the release of extremely efficient AI models that can compete with cutting-edge products from US companies such as OpenAI and Anthropic.

Founded in 2023, DeepSeek has achieved its results[2] with a fraction of the cash and computing power of its competitors.

DeepSeek’s “reasoning” R1 model, released last week, provoked excitement among researchers, shock among investors, and responses from AI heavyweights. The company followed up on January 28 with a model[3] that can work with images as well as text.

So what has DeepSeek done, and how did it do it?

What DeepSeek did

In December, DeepSeek released its V3 model[4]. This is a very powerful “standard” large language model that performs at a similar level to OpenAI’s GPT-4o and Anthropic’s Claude 3.5.

While these models are prone to errors and sometimes make up their own facts[5], they can carry out tasks such as answering questions, writing essays and generating computer code. On some tests[6] of problem-solving and mathematical reasoning, they score better than the average human.

V3 was trained at a reported cost[7] of about US$5.58 million. This is dramatically cheaper than GPT-4, for example, which cost more than US$100 million[8] to develop.

DeepSeek also claims to have trained V3 using around 2,000 specialised computer chips, specifically H800 GPUs made by NVIDIA[9]. This is again much fewer than other companies, which may have used up to 16,000[10] of the more powerful H100 chips.

On January 20, DeepSeek released another model, called R1[11]. This is a so-called “reasoning” model, which tries to work through complex problems step by step. These models seem to be better at many tasks that require context and have multiple interrelated parts, such as reading comprehension and strategic planning.

The R1 model is a tweaked version of V3, modified with a technique called reinforcement learning. R1 appears to work at a similar level to OpenAI’s o1[12], released last year.

DeepSeek also used the same technique to make “reasoning” versions of small open-source models that can run on home computers.

This release has sparked a huge surge of interest in DeepSeek, driving up the popularity of its V3-powered chatbot app[13] and triggering a massive price crash[14] in tech stocks as investors re-evaluate the AI industry. At the time of writing, chipmaker NVIDIA has lost around US$600 billion[15] in value.

How DeepSeek did it

DeepSeek’s breakthroughs have been in achieving greater efficiency: getting good results with fewer resources. In particular, DeepSeek’s developers have pioneered two techniques that may be adopted by AI researchers more broadly.

The first has to do with a mathematical idea called “sparsity”. AI models have a lot of parameters that determine their responses to inputs (V3 has around 671 billion), but only a small fraction of these parameters is used for any given input.

However, predicting which parameters will be needed isn’t easy. DeepSeek used a new technique to do this, and then trained only those parameters. As a result, its models needed far less training than a conventional approach.

The other trick has to do with how V3 stores information in computer memory. DeepSeek has found a clever way to compress the relevant data, so it is easier to store and access quickly.

Computer screen with deepseek logo and the words into the unknown underneath it.
DeepSeek has shaken up the multi-billion dollar AI industry. Robert Way/Shutterstock[16]

What it means

DeepSeek’s models and techniques have been released under the free MIT License[17], which means anyone can download and modify them.

While this may be bad news for some AI companies – whose profits might be eroded by the existence of freely available, powerful models – it is great news for the broader AI research community.

At present, a lot of AI research requires access to enormous amounts of computing resources. Researchers like myself who are based at universities (or anywhere except large tech companies) have had limited ability to carry out tests and experiments.

More efficient models and techniques change the situation. Experimentation and development may now be significantly easier for us.

For consumers, access to AI may also become cheaper. More AI models may be run on users’ own devices, such as laptops or phones, rather than running “in the cloud” for a subscription fee.

For researchers who already have a lot of resources, more efficiency may have less of an effect. It is unclear whether DeepSeek’s approach will help to make models with better performance overall, or simply models that are more efficient.

References

  1. ^ shockwaves through the tech community (www.theverge.com)
  2. ^ achieved its results (www.technologyreview.com)
  3. ^ a model (techcrunch.com)
  4. ^ V3 model (arxiv.org)
  5. ^ sometimes make up their own facts (arxiv.org)
  6. ^ some tests (www.anthropic.com)
  7. ^ reported cost (www.scmp.com)
  8. ^ more than US$100 million (www.wired.com)
  9. ^ H800 GPUs made by NVIDIA (www.reuters.com)
  10. ^ up to 16,000 (www.nytimes.com)
  11. ^ called R1 (arxiv.org)
  12. ^ OpenAI’s o1 (openai.com)
  13. ^ V3-powered chatbot app (www.theguardian.com)
  14. ^ massive price crash (www.ft.com)
  15. ^ has lost around US$600 billion (www.abc.net.au)
  16. ^ Robert Way/Shutterstock (www.shutterstock.com)
  17. ^ MIT License (opensource.org)

Read more https://theconversation.com/deepseek-how-a-small-chinese-ai-company-is-shaking-up-us-tech-heavyweights-248434

Times Magazine

Building a Strong Online Presence with Katoomba Web Design

Katoomba web design is more than just creating a website that looks good—it’s about building an onli...

September Sunset Polo

International Polo Tour To Bridge Historic Sport, Life-Changing Philanthropy, and Breath-Taking Beau...

5 Ways Microsoft Fabric Simplifies Your Data Analytics Workflow

In today's data-driven world, businesses are constantly seeking ways to streamline their data anal...

7 Questions to Ask Before You Sign IT Support Companies in Sydney

Choosing an IT partner can feel like buying an insurance policy you hope you never need. The right c...

Choosing the Right Legal Aid Lawyer in Sutherland Shire: Key Considerations

Legal aid services play an essential role in ensuring access to justice for all. For people in t...

Watercolor vs. Oil vs. Digital: Which Medium Fits Your Pet's Personality?

When it comes to immortalizing your pet’s unique personality in art, choosing the right medium is ...

The Times Features

The Role of Spinal Physiotherapy in Recovery and Long-Term Wellbeing

Back pain and spinal conditions are among the most common reasons people seek medical support, oft...

Italian Lamb Ragu Recipe: The Best Ragù di Agnello for Pasta

Ciao! It’s Friday night, and the weekend is calling for a little Italian magic. What’s better than t...

It’s OK to use paracetamol in pregnancy. Here’s what the science says about the link with autism

United States President Donald Trump has urged pregnant women[1] to avoid paracetamol except in ...

How much money do you need to be happy? Here’s what the research says

Over the next decade, Elon Musk could become the world’s first trillionaire[1]. The Tesla board ...

NSW has a new fashion sector strategy – but a sustainable industry needs a federally legislated response

The New South Wales government recently announced the launch of the NSW Fashion Sector Strategy...

From Garden to Gift: Why Roses Make the Perfect Present

Think back to the last time you gave or received flowers. Chances are, roses were part of the bunch...

Do I have insomnia? 5 reasons why you might not

Even a single night of sleep trouble can feel distressing and lonely. You toss and turn, stare...

Wedding Photography Trends You Need to Know (Before You Regret Your Album)

Your wedding album should be a timeless keepsake, not something you cringe at years later. Trends ma...

Can you say no to your doctor using an AI scribe?

Doctors’ offices were once private. But increasingly, artificial intelligence (AI) scribes (al...