Meanwhile, Anthropic have just released a plugin for Claude Cowork to do equity research.
https://claude.com/plugins/equity-research
I wasn't going to comment on this article, but now I can't resist. Firstly, i didint finish it, id had enough. i think it says more about the state of skittishness (or FUD as SM would say) in the market that this article got so much airtime than anything else. as ive written here before with Ai there is endless scenarios you can paint, and this is one of the darkest. IMO the stuff served up here ranges from possible to fantasy. the equivalent of going to a beach that just suffered a horrendous shark attack and yelling "shark" and watching the ensuing chaos..
People can do whatever they choose (and they will), i choose to listen carefully to the main players and see how the narrative changes. Take the Dwarkest podcast with Dario recently. Now he has his $30B in, he started backpeddling a little talked about the risks of timing demand and the lack of diffusion of AI into the economy. we shall see
I have to admit I’ve had half a bottle of wine so am presuming I’m a bit more susceptible to an argument than usual. Still, there was very little in this (very long) narrative that didn’t seem plausible. Perhaps the timeline is accelerated, but other than that it is dangerously realistic
Have to read it again without the aid of the Argentinian Malbec
Well worth the time
Like everyone else I’ve been thinking a lot about AI disruption lately.
Keen to get input from the Strawman intelligentsia on this too, so please share any insights / rebuttals!
I’ve tried to be systematic about this and have split my pursuits into 2 areas – 1) how I can and should be using it personally and professionally – including to Equity Research and 2) how it will likely impact businesses and therefore investing over the LT.
On point 2, I have tried to go back to first principles and build an understanding of the underlying Economics of AI.
I’m still very much learning about all of this so putting thoughts on paper here more than anything else…
Solutions vs Systems
One framework suggests there are 2 aspects to how AI Affects the economics of a business and its products. Firstly Solutions, secondly Systems.
For Solutions, a business can (will need to) apply AI capabilities to its products and how it delivers these to customers to make them faster, more effective, more customisable by end users, etc – this will mean making and delivering them with less human intervention and therefore less cost.
For Systems, a business will put AI at the centre of how its operates and the products it develops and how these are delivered. This is a simple but very important distinction which comes down to the economics of AI.
What starting with AI really allows you to do is to decouple prediction from judgement in decision making. This is likely too hard in existing organisations where these are bundled in people, process, departments, all the way up to the CEO. Much better to start with an AI centric structure and as the cost of prediction goes to zero and speed and accuracy improve massively, add judgement on top to arrive at better decisions, better products and an organisation that is leaner, more scalable and cheaper to run.
The Innovator's Dilemma
I was thinking how this could apply to PME (same goes for WTC). PME is the leader in their field by some distance, they are AI enabled and use it extensively so seem well placed to capitalise on the opportunity and defend against the treats from AI.
However, they are still vulnerable to Clayton Christensen’s The Innovator's Dilemma (1997) This explains why leading companies fail by doing "everything right"—listening to customers and focusing on profitable, high-end products. These firms neglect disruptive, low-end innovations that eventually improve, steal market share, and displace them. Successful firms must embrace, rather than ignore, these disruptive technologies.
In this case AI native businesses / products can in theory quickly and cheaply service customers at the lower end of the food chain with a cheaper, low spec product that is good enough for them. Over time, they add capabilities (and collect data) so that they can more effectively compete with the larger incumbents. AI should speed up this Innovator's Dilemma process – potentially a lot depending on how well the incumbent’s economic moat holds.
Some Examples
Here PME (or WTC) could spin up an AI native department / division / business unit to devise AI centric products to segment the market (become the Jetstar to their Qantas). This would compete head to head with any new AI startups / upstarts at the cheaper end of the market and should do well given their existing internal data set and industry expertise. It also leaves the legacy, premium, high cost, high spec, highly profitable Visage (Cargowise) business in tact meaning they don’t need to disrupt themselves too much too fast. This would also provide valuable lessons on how to evolve (or rebuild) the legacy business with emerging AI tech over time.
Same goes for TNE.
Same goes for … KYP?
See separate post under KYP thread here for more - https://strawman.com/forums/topic/12714#post-41983