1/ Happy Monday Twitter. ⚠️ #tweetstorm 🌪️ #alert ⚠️
My #LegalGeek #FOMO peaked when I missed @nwaisb talk on #AI #hype which is still ALL OVER MY TWITTER FEED. Plus now I have #ACCAM18 #FOMO so here's a bunch of 🤔re #AI, #MachineLearning, #EmergingTech #hype, and #marketing.
My #LegalGeek #FOMO peaked when I missed @nwaisb talk on #AI #hype which is still ALL OVER MY TWITTER FEED. Plus now I have #ACCAM18 #FOMO so here's a bunch of 🤔re #AI, #MachineLearning, #EmergingTech #hype, and #marketing.
2/ As @ZachAbramowitz thinks there's no fun in agreeing: no, I don't think @nwaisb / @KiraSystems & others have benefited (on net) from the AI hype. I'd say hype hurts us all, bc it trades short term gainz to a few players for long-term pain shared by all.
5/ Esp for #EmergingTech that probably requires new thinking around process and talent, prospects have to be qualified for readiness to buy & deploy. Converting leads thru the sales cycle costs startups time + effort and ultimately can lead to higher cost of customer acquisition.
7/ So I agree w @nwaisb that #hype is the scourge. BUT (a) I am not sure truth vs falsity is the optimal lens here + (b) I wouldn't put this entirely on tech press. I'd actually ask the startups (incumbents too): how effective is your #marketing? how good is your #storytelling?
8/ Sure, nobody holds a monopoly on truth, but what is "truth" in this regard? AI is a thing that exists? Company X uses it? Both conditions can be "true" and still result in a product that sucks bc it doesn't create gain or relieve pain for the customer.
9/ I think some of the debate over AI hype vs. "legit" AI is intrinsically product-centric, not customer-centric. (Back to that in a hot minute)
Staying with product claims, I agree 💯 with @edsohn80 here bc he speaks so much sense:
Staying with product claims, I agree 💯 with @edsohn80 here bc he speaks so much sense:
10/ YES, this stuff has been rendered MEANINGLESS - but "AI" is esp tricky bc common usage has gone off the rails and is now a very loose term wrapped around multiple conceptual or technical categories & appears to refer to moving goalposts ("once it works, it's just software").
11/ When it comes to assessing technical claims, I think more precise terminology is superior to vague buzzwords bc we can draw practical distinctions + abstract out the use case implications: rules-based vs ML, explainable AI vs genetic algorithms, current limits of NLP, etc.
12/ BUT this segues into broader Qs: how sophisticated do buyers need to be to buy AI tech + who will bear the costs of educating the market? I think in the short-term, most buyers are still unsophisticated. But there is better content marketing now trying to explain HOW AI WORKS
13/ Still lots of sloppy language and lazy thinking. I tend to read hype as incompetent marketing with possible signals to bad product strategy. Instead of true/false, real/hype - I'd think startups would benefit more from the improved ability to explain HOW OUR PRODUCT HELPS YOU
14/ And sometimes, that value prop is still getting baked, in POCs + PAYG arrangements. So for that reason, I'm not sure I totally agree with @nwaisb in dismissing those - still newsworthy IMO. (Plus, good competitive intel for all market players if you can find signal in noise.)
15/ Legal tech press is overloaded and under-resourced. I appreciate this - i spend a lot of time on research, legwork, analysis, thinking, etc. to generate content. BUT I tend to think there are a few easy gut checks to test maturity, adoption & scale (and we need more of this).
16/ Here is a great + accessible piece by @jtashea on common pitfalls around AI "definitions" (which he says are "intentionally broad and squishy, making their utility limited" - YAAS) and gives sound advice on how to think and talk about AI. abajournal.com/lawscribbler/a…
