Blain’s Morning Porridge Jan 28^th 2025: Pens vs Pencils – What was AI all About Anyway?

“Urg.. that onion tasted funny.” … “That’s because it was a 1 billion guilder tulip.”

It’s going to take some while for markets to figure out the consequences and implications of DeepSeek on the AI narrative, including how much false value has been factored into US stocks in particular. What will be the next no see-um remains the eternal question for markets!

I’m peeved with myself for not using another story from the space age yesterday when I wrote about DeepSeek. While everyone else is calling the dramatic launch of the Chinese AI firm The Sputnik Moment,  I’m thinking about the apocryphal story of how NASA spent billions developing a zero-gravity pen, while the Russians used ….. pencils.

Pens vs Pencils might be the story of how AI changes the modern economy. Being able to scribble down fuel/weight orbit equations while life-boating Apollo 13 back to Earth was a critical moment, but a footnote in the remarkable history of how going to the moon changed technology. Did the billions spent on developing a space pen really change the story? Nasa spent even more developing the most complex machine in human-history, the Space Shuttle (with 2 tragic outcomes) which never hit anywhere close to the expected launch cadence, while the Soviets are still safely launching the same old Soyuz/Proton capsule/booster combination to the ISS decades later.

The DeepSeek tremblor has been an eye-opener. Tech markets took a spanking yesterday as they tried to cope with the implications for growth across every aspect of the developing AI community, from Chips, logistics, and Energy. Nvidia experienced the largest valuation fall in history. Difficult questions are being asked about market direction. Naturally, there are some folk saying “buy the dip” or “this is a buying opportunity”, but those might be seeking an exit, or they are delusional true-believers. Even if DeepSeek proves less than what it seems (a distinct possibility), the market has had a valid wake-up-and-smell-the-coffee moment.

We should always pay attention to reality checks. The questions this morning are wide ranging; will AI really impact humanity as much as promised, and make the world a better place? It should be part of the solution, another step in the growth path of mankind, but… to be honest, it feels like the potential of AI was hi-jacked by the financial markets from the get-go! Every banker and tech VC funder found an axe on how to make money. AI has become a label to pin visions of fantastical wealth upside upon.

In the naked pursuit of wealth, what has AI actually spawned?

• Nvidia (briefly) as the world’s most valuable firm as the picks and shovels of a new industrial revolution? Chips are a very competitive space. DeepSeek’s success has apparently come on the back of understanding how to use older chips better.
• A host of AI based start-up unicorns? The Dot.Com crunch – a quarter century ago seems to have been forgotten… (Bankers who remember it, are now old and ignored.. out of touch..)
• The billions spent by the hyperscalers like Meta, Alphbet and Amazon to redefine, reset and relaunch their aging businesses as fresh, new, AI based mega-firms? Easy money for over a decade and untaxed corporate wealth enabled firms that should have been replaced by new entrants to survive far longer than financial evolution ever intended.

What has actually been achieved by AI?

• We have the annoyance of Microsoft’s AI co-pilot widget-thingy trying to tell you how to write articles and emails.
• We can ask ChatGPT, Grok, or now DeepSeek, to do our thinking for us and answer whatever they think the question is? (The real risk of AI is not that it wipes it out, but we forget how to think for ourselves…)
• Or, the repetitive drone of already rich tech entrepreneurs telling us yet again how AI is going to change the world?

At this point, folk will be crying foul: there are valid uses of AI in the diagnosis, interpretation, and actions regarding data sets: give AI enough data and it will determine potential outcomes far more reliably than humans. From drugs to the environment – to filing our taxes on time, AI can probably help. Yet acting upon AI will prove sticky.

—-

An aside: A number of years ago I was quite sick. I got friendly with the surgeon who put me back together again. He had a vision for the UK’s National Health Service, where all the data held on every UK citizen would be digitised, collated and used to predict the most likely diagnoses, the apt treatments, and routes to best outcomes for patients. Preventive medicine would be boosted by yearly blood tests to spot changes and threats coming down the line at us. All that data properly used would be a health game changer.

Yet, that vision of an AI-enabled NHS is still nowhere on the agenda – the reality is repeated database project failures, day-to-day budgeting (where long-term is tomorrow), and too many mangers of failing trusts and vested interests in health ensure it won’t change despite the dedication of the medical staff to treatment.

—–

Implementing AI will be as difficult as any corporate or national turnaround. It doesn’t help that AI has now become a “label” to hide behind which ascribes value. It is used by unscrupulous types to create unsustainable narratives. Despite the emergence of AI, the most successful firms are still those with the best sounding narratives.

Tesla is a great example. Many years ago it was an electric car company, trading on a PE multiple 10x that of any other auto-company because it dominated the then new EV sector and everyone was going to buy an EV. It was then a battery company. Then it was self-driving cars – which Musk has been promising tomorrow, but will likely never fully deliver. (Telsa’s Autopilot tech lags Waymo in busy urban environments and is still far from safe self-driving.) Last year, Elon decided Tesla was part of his new AI business (which has been kicked into second-tier irrelevance by the announcement of Stargate last week and DeepSeek this week).

Now Tesla is about selling Robo-taxis… apparently. Yesterday I was talking to a chum who values Tesla at $15 trillion because of what Musk is going to produce in Humanoid Robots.. FFS… Musk dressed a guy in a Robot Suit when he first tested that axe on his ongoing “Technoking” following.” To be blunt Tesla is worth so much because the market is waiting to see how he benefits from being “First Buddy” to Trump, and he understands the power of media.

Since its inception Tesla has never been valued on what it is, but on what Elon tells us it is going to be.. Should we be listening to him? If I was going to be blunt, its a niche producer of electric vehicles which are about to be clobbered by Trump removing subsidies, by global competition and tit-for-tat tariffs, and by the fact the liberals that supported EVs now detest Musk.

But Tesla is just one of the multiple unsustainable narratives the market seems to believe are set-in-stone truths:

Nvidia is the most valuable company in the world because only it can make the chips to fuel the AI revolution? Not so, and who needs Nvidia chips to make it happen. The revolution will not be televised. Nvidia is worth so much because the market bought into a narrative that needed scarcity to make it exciting.

Apple is set to remain the most valuable brand on the planet. Although it can’t get its lacklustre AI offering to excite anyone, Apple’s legions of sophisticated buyers will keep buying the Apple ecosystem and iPhones.. right up to the moment they don’t. I would simply say.. Apple Stores: don’t they look tired?

The hyperscalers – what a wonderful word – Amazon, Microsoft, Meta and Alphabet are the ones potentially left looking most foolish by DeepSeek. They are planning to spend gazillions of dollars on new chips, data-centres, and even their own nuclear power! They expect their customers to upgrade accordingly. I can’t help but recall how Facebook changed its name to Meta to reflect how we would all transact and live in the Metaverse… where now pixellated tumbleweed blows down the pixelated empty lots…

One real factor that makes economies grow is energy. The thinking was the amount of power it would require to make it happen, to power the chips, the data-centres, the models, the cloud, the internet, etc.. was a competitive advantage for the USA. Suddenly we find someone has done it for pennies, using repurposed gaming chips plugged into the local grid? Anyone can play. That is a narrative changer.

Oh dear..

Let’s not be overly negative about the AI dreamscape…. The world should be about evolution, change, growth and how new ideas change it for the better. But, yet again, it’s been distracted by greed narratives – and it’s probably not a bad thing to give it a sharp reality slap every now and then.

I suspect AI will remain the market’s tender spot through the year…

Out of time, and back to the day job..

Bill Blain

Author of the Morning Porridge

www.windshift.capital

billblain@morningporridge.com