In addition to this newsletter, I recommend some other great ones. All free. Check them out here.
Friends,
I touched on DeepSeek last week, but it feels like it’s changed the world over the course of a week, so I wanted to point you to resources that helped me get my head around it. Top to bottom, from surface understanding to seeking deeper (see what I did there?):
DeepSeek R-1 Explained A no-nonsense FAQ (for everyone drowning in DeepSeek headlines)
DeepSeek’s Rise: How a Chinese Start-Up Went From Stock Trader to A.I. Star
The DeepSeek app is impressively strange. Here’s what the fuss is about
Meet Alibaba's Qwen 2.5, an AI model claiming to beat both DeepSeek and OpenAI's ChatGPT
DeepSeek Debates: Chinese Leadership On Cost, True Training Cost, Closed Model Margin Impacts
I only deep-dived into Jevon’s paradox when Satya Nadella mentioned it this week in the context of how DeepSeek had changed the efficincy level of AI. (Here’s the wiki, he linked to)
Jevons paradox comes from how steam engines changed coal consumption in the 19th century. It seemed counterintuitive at the time. Engineers kept making the engines more efficient, yet coal consumption rose rather than fell. That pattern could appear strange, though it is logical. That technology became more accessible, so more people used it, driving consumption up. Pretty obvious.
The same principle shows up whenever we make tools more powerful and affordable. Deepseek’s launch highlights this pattern in AI. The system promises greater efficiency by saving time, cutting down on redundant tasks and speeding up the search for insights. The engineering is impressive, and many early adopters find it transforms their work. They finish their data analysis in half the time and at a fraction of the cost they expected. Some say they’ve never seen anything like it. And that might actually be the effect of users returning to AI, after only experiencing it through Chat GPT’s launch two years ago. Either way, it’s a sort of hailo effect.
After improvements like this, a second wave of adoption kicks in - which is what we are seeing now. The system encourages more queries, more experiments and more data-crunching than ever before. Lower costs per query often mean teams run more queries. Teams that previously ignored data analytics begin to see it as a practical investment, because they can glean insights in minutes. That’s the essence of Jevons paradox. We make a tool less resource-intensive per use, but we use it more often, so resource consumption climbs. We also make it more marketable. DeepSeek is a great name, and it’s launch at a peak AI moment (Stargate) propelled this Chinese startup to the front page of every new outlet as it pummelled the shares of nearly every AI-relevant tech company.
In AI, the resources used are not only in the direct costs of queries. Data centres run on electricity and produce heat. In 2022, some estimates suggested that data centres represented around one percent of global electricity consumption. As the world accelerates AI adoption, some analysts predict an even larger share of energy use. Chipmakers keep improving hardware, but Jevons paradox implies that this efficiency can lead to greater overall resource demand. More training runs, more model deployments and more usage feed into that growth cycle.
In practice, many users focus on immediate goals like shipping a product faster or beating a deadline. The paradox emerges as an unintended consequence. Efficiency invites heavier usage, which can negate part of the original benefit if people don’t keep track of usage spikes. On the other hand, organisations find new insights and new revenue streams, so the total impact isn’t purely negative. It’s a trade-off that depends on awareness and self-restraint.
Jevons paradox isn’t a universal law that dooms every efficient technology to an explosion in resource use. It depends on human behaviour and market forces. Deepseek’s launch shows how exciting these tools can be when they cut the cost of analysis, but it also reminds us that soaring efficiency can create extra demand. That might lead to a larger energy footprint and higher levels of consumption overall. The paradox resides in the gap between saving resources in a single task and embracing more tasks in the long run. Each new wave of AI innovation makes this conversation more urgent.
Deepseek’s launch shows traits of Jevons paradox but isn’t appear to be a perfect example. The paradox occurs when efficiency gains drive overall consumption higher, erasing the initial savings. In one week, while Deepseek has enabled teams to run more analyses and take on bigger projects, leading to higher data costs, many companies still operate within budgets and prioritise value-driven tasks. This limits uncontrolled expansion. While Deepseek changes behaviour by making data analysis easier, the increase in usage doesn’t necessarily cancel out efficiency gains. The extent to which it follows Jevons paradox depends on whether rising demand outpaces cost savings in the long run. 2025 will prove that out as we see this arms race continue.
Stay Curious - and don’t forget to be amazing,
Here are my recommendations for this week:
One of the best tools to provide excellent reading and articles for your week is Refind. It’s a great tool for keeping ahead with “brain food” relevant to you and providing serendipity for some excellent articles that you may have missed. You can dip in and sign up for weekly, daily or something in between -what’s guaranteed is that the algorithm sends you only the best articles in your chosen area. It’s also free. Highly recommended Sign up.
Now
What If Humans Don’t Outperform AI In ‘Human Skills’? -Many have predicted that as AI improves, it will commoditize technical skills and knowledge, but accentuate the things that make us human—things like our empathy and connection with other human beings. Or our ability to communicate.
How to Remain a Reality-Based Human in 2025 - If the first month of 2025 with its blazing fires, bitter cold snaps, AI Armageddon and the pilot episode of “Trump 2” is any indication, the coming year is bound to challenge anyone trying to hold fast to the norms of basic human existence. But however broken your brain may feel, there’s no need to pass the steering wheel entirely to GPS.
How to maintain a healthy gut - With a few lifestyle and dietary changes, you can protect your gut microbiome, boost your immunity and improve your mood
The Strange Power of Laughter - An anthropologist explores laughter as a far more complex phenomenon than simple delight—reflecting on its surprising power to disturb and disrupt.
Bill Gates Isn’t Like Those Other Tech Billionaires - The Microsoft co-founder and philanthropist is taking a break from the future to examine his past — and mulling where the billionaires now fit in.
Next
The anatomy of a bubble bursting - When lots of people are worried about bubble valuations in stocks or a specific sector, all it takes is a small poke to make the whole thing wobble precariously.
No one is disrupting banks - Perhaps, disrupting big banks takes more time or even a change in generations of consumers. Or, maybe, banks and Fintech companies will need to learn how to co-exist. In any case, I don’t think the future of the Fintech industry will be boring.
New glowing molecule, invented by AI, would have taken 500 million years to evolve in nature - An artificial intelligence model has created a new protein that researchers say would have taken 500 million years to evolve in nature — if nature were capable of producing such a thing.
2025 will likely be another brutal year of failed startups, data suggests - More startups shut down in 2024 than the year prior, according to multiple sources, and that’s not really a surprise considering the insane number of companies that were funded in the crazy days of 2020 and 2021. It appears we’re not nearly done, and 2025 could be another brutal year of startups shutting down.
AI Needs a Lot of Computing Power. Is a Market for ‘Compute’ the Next Big Thing? Startup trading platforms seek to capitalize on AI’s need for vast amounts of processing capacity
Free your newsletters from the inbox: Meco is a distraction-free space for reading newsletters outside the inbox. The app has features designed to supercharge your learnings from your favourite writers. Become a more productive reader and cut out the noise with Meco - try the app today
If you enjoyed this edition of Box of Amazing, please forward this to a friend. If you share it on LinkedIn, please tag me so that I can thank you.
If you are interested in advertising in this newsletter, please see this link
Definitely have some research to do !!