Vertical AI = Expert System, just an updated rebranding from 1980s to now. An expert system may be defined as:<p>>a piece of software which uses databases of expert knowledge to offer advice or make decisions.<p>The methods of inference have improved from predicate logic to statistics and "machine learning" now that computers have gotten much faster.<p>(I'm bootstrapping an expert system myself)
Hands down one of the best articles on the topic I've read - so many "wish I had known this years ago" types of insights.<p>The video contains bits that aren't fully reflected in the post itself, so it's worthwhile to watch the whole thing. The most important part IMHO is that you absolutely have to start with a known problem and work out the solution from there, not vice versa. This is a trivial insight for most startups, but a surprisingly common trap for even the smartest AI/ML people.
It's very easy to get blinded by the sheer awesomeness of the data/model you stumbled upon, all while ignoring that you haven't fully grasped the problem space yet.
My company[1] Analytical Flavor Systems is a vertically integrated domain-expert based AI for new product development, flavor profile optimization, and predictive manufacturing in the food and beverage industry.<p>Like the article and some of the comments here suggest, it took years of domain expertise (most of the team comes out of the Tea Institute at Penn State, a research Institute for tea and tea tasting), followed by years of R&D to collect the proprietary data-sets and develop the models. And then it took a year or so to build a product around the AI's predictive capabilities - this isn't the shortest or easiest path, but we're still going strong!<p>I think companies like this are hard to build, hard to fund, and hard to compete with.<p>Where I disagree with other comments is on the competitive side; we've developed a few of our own algorithms[2] (not generic or even "played with some options" neural nets / deep learning) trained on specialized and proprietary data set from years of work and collection - now that we've dug our moat, I don't think anyone will be competitive with out specialized AI for modeling human sensory perception and predicting preferences[3] of food and beverage products anytime soon!<p>[1] www.Gastrograph.com<p>[2] <a href="https://gastrograph.com/resources/whitepapers/local-fisher-discriminant-analysis-on-beer-style-clustering.html" rel="nofollow">https://gastrograph.com/resources/whitepapers/local-fisher-d...</a><p>[3][PDF] <a href="https://gastrograph.com/resources/whitepapers/2017-market-preference-tasting-panels.pdf" rel="nofollow">https://gastrograph.com/resources/whitepapers/2017-market-pr...</a>
AI is not like ordering your lunch : "I don't like to sour, can you make it a bit spicy", doesn't work for AI. Cutting edge AI solutions require highly customized frameworks. The number of ingredients involved are very large in number and you can't go on making on creating a list of parameters to create your own "recipe". Therefore, a platform or stack will only affect solutions which are have least variance across industries - language translation, object recognition, photo tagging etc.<p>Industry specific customized solutions would require SMEs to build solutions. At this point, you are essentially buying people skills and not products. You'd also see a large increase in price points here which companies may not be willing to pay. Everyone wants something quick and cheap. AI is not a panacea or a magic wand. The hype is more detrimental to the progress of AI. Once the disillusionment sets in - people will blame the technology rather than people who took decisions to use AI in the wrong contexts.
What's the opposite of Vertical AI? Is it a monolithic chunk of code like an operating system with millions of lines?<p>I've always been a bit confused about what the end goal will look like for 'general AI'.<p>Or is there potential for some general platform that these vertical AI systems can plug into, similar to the app store, but with some type of cross vertical communication so they be layered on top of each other.
What I would like to understand is why are these startups defensible.<p>The tech value proposition is in running these algorithms, often tensorflow at scale cheaply in production.
Companies like Google/Facebook/Palantir often have access to very similar supposedly hard to get to datasets, plus a lot more engineering expertise to running these systems at scale.<p>Why can't they start playing whack-a-mole pumping out vertical products presenting a serious threat to these smaller startups. Maybe it's not worth it for them but there is a fair bit of cash there ?<p>For example Deepmind with healthcare, and the google jobs API ?
As someone who is a subject matter expert that is using more and more AI/ML for their research, this makes intuitive sense. I could keep up with current AI research, but I would barely be treading water and couldn't create anything of substance.<p>However, incorporating mature ML methodologies (meaning it has a library) to subject-matter problems is now adding a tremendous amount of value in the research I do.
Since a long time, I try to figure out an AI-based vertical in the tech recruiting space. However, I am stuck in agency mode (<a href="http://coderfit.com" rel="nofollow">http://coderfit.com</a>). If you have good ideas how AI can help to source and hire software engineers, I have done the tedious work of getting paying clients and a database with engineers who are looking for a job.
In my opinion the mainstream AI startups are described in article, mainstream AI startup = unique data + algorithms. I think it is good to not be in AI startup mainstream and work on things that maybe won't make you rich but will make you happy.
If you can deliver 'a totally new opportunity through rich domain modelling' (ie. You have a solution to a problem that hasn't been solved before), and your solution:<p>- Is fundamentally tied to data that is proprietary and difficult to gather.<p>- Is intrinsically extremely complex to build and maintain.<p>- Can only be built by a diverse team which is difficult to gather.<p>Then... well, yes, I guess you could say that those of good metric for determining if a product can easily <i>be replicated</i>.<p>...but that's not the same thing as it being <i>useful</i> or <i>profitable</i>.<p>It just means that the team has a bit more time to try to figure out those other two important things before someone else comes along and copies what they've done.<p>> My claim is that Vertical AI startups are inherently defensible.<p>Putting 'vertical' in front of 'AI' doesn't magically make things better than just 'AI'.<p>The problem with these products is you can't take a trivial 'proof of concept' or MVP, pitch it and then roll it out 'into production'.<p>This isn't some 'smoke and mirrors' jazzy demo of a website & app combo you can go away and implement properly later... the proof of concept you build may not scale. Like... it may actually <i>not be possible</i> to scale. Maybe it takes too much compute to train; maybe it takes data you don't have; maybe it turns out your data isn't suitable.<p>You think building a business and getting users and sorting your workers and so on isn't hard enough?<p>Try adding a product that may or may not <i>actually ever work</i> into the mix. Sound scary yet? It should sound scary. That's the sound of money draining out of a hole in the floor.<p>...and sure, you argue, these models <i>do work</i> and they <i>are</i> good; but there's the catch right there... :)<p>...if you use a model that <i>does</i> work and isn't risky and does use available data... then you lose all the points at the top that made it an interesting business to invest in.