@chad987, great discussion, I like your comments, it is obvious that you are thinking deeply about all of this.
Regarding your first point that Facebook's & Google's advertising revenues are at record levels and that "Mark still claims they are hit by Covid"... I think what Mark is claiming is that the internal AI data engineers at FB/Google have reprioritised their resources into new AI products, away from advertising optimisation projects.
Mark is not suggesting that advertising revenue for FB/Google isn't growing (they are blowing the lights out!) but rather that FB/Google's
spend on optimising advertising AI (via new data labelling projects) shifted to other AI projects (partly due to a realisation that they have an over-reliance on advertising revenue). This is what hit Appen's top line.
Regarding the automotive market, I don't think this is the exclusive domain of Scale AI. From what I know, Scale AI work with GM & Toyota and maybe one other major. Whereas
Appen works with 7 of the 10 top auto companies and tier 1 suppliers. Appen retains a significantly larger scale compared to their competitors in this industry (as in most industries). I think Appen could take a very large market share in this market, particularly in the Chinese automotive market, where Scale AI doesn't have a presence at all (yet).
"What I see is the mega tech switching more to weak supervised/transfer learning, and even unsupervised learning." That is probably a fair assessment Chad, but the industry will still require labelled data in conjunction with these methods in the medium term.
"If unsupervised learning becomes the real deal in 10 years time, Appen's revenue will begin to drop well before then, with most of it gone in year 10. now, how many years of DCF is required for the current forward p/e of ~33? (FNarena estimate ~0.35c eps this year). Appen should be priced as a short duration miner like iron/gold, not a long duration tech growth stock." — This is a really good comment Chad and cuts into the heart of why Appen has never been priced as exorbitantly as its WAAAX peers.
The truth is that it is hard to value the long-dated cash flows of the business with high levels of certainty. However, some believe that we are only at the very early stages of the AI lifecycle and therein lies the debate. Also, Appen's valuation relies on how well they position themselves to pivot into adjacent opportunities across that decade.
Appen has the opportunity to leverage a seeming commodity — data labelling — as a wedge into a key position in the rapidly growing and evolving AI and ML space. By building infrastructure, Appen has a chance to help grow AI/ML usage from only 8% of companies today, to all companies within the next two decades. We are at the beginning of the S curve.
In a decade, if Appen is successful, any company that wants to build something using AI or ML will just stitch together six different Appen services like they stitch together AWS services to build something online today. Appen could collect or generate data for you, label it, train the machine learning model, test it, tell you when there’s a problem, continue to feed it fresh, label data, and on and on.
Appen's ambitions are obfuscated by its starting point: using humans to build a seemingly commodity product. A bet on Appen is a bet that data labelling is the right starting point to deliver the entire suite of AI infrastructure products... And if so, the current valuation looks tiny indeed...
(20min delay)