Incredible thoughts #59
another newsletter, not about things, but about other things that the things made me think about
Sometimes I read something that I just really enjoy. I probably can’t put my finger on why, but some things just really land for me.
I read something like the other weekend, and just loved it. In some ways, articles like these make me think maybe AI won’t put creatives out of work. Of course, I know that it’s just a series of words, so why couldn’t AI put it together? Maybe it just comes back to the value we attribute to the human authors and their life experience and world view from which they are creating.
Anyway, on with the week:
🅰️ 🅱️ A/B testing
⚖️ 📐 Legal tech and product market fit
🤝 📈 Denton’s CTO @ Dentons on the Lexi/Harvey deal
📼 🔍 Round and round we go - YouTube’s AI search
💅 👁️ Beauty maybe isn’t in the eye of the beholder?
🙋♂️ 🤪 Other stuff I found interesting
The $170billion A/B test
Yep, the piece that I really got excited about was a long-form piece on travel website booking.com’s approach to A/B testing. Wild.
There are so many interesting bits of information buried away, so many insights and food for thought. Maybe part of the reason I loved this piece quite so much was that its an example of people following a fairly standard growth playbook to the letter, and it working (which I’d say isn’t the norm):
Sounds easy on paper, but if it was that easy, of course nobody would fail.
I particularly liked the acknowledgement of the fact that the site had effectively been designed and built by the customers:
Legal AI & Product Market Fit
Another thing I’ve been thinking about recently, on the back of the recent FT and BI articles, as well as the flurry of activity in the legal-tech market, is legal-AI and Product Market Fit.
I don’t think it’s a controversial take to suggest that AI hasn’t found product market fit in legal yet. I know we have legal-tech unicorns and some big-money valuations, but have they found product fit yet? It feels as though some of the current legal AI vendors have the resources, capital, tech and brand to build most things, but it feels like they're still trying to work out what the perfect solution is.
I saw one quote, which whilst not at all the only thing I think of when I think of PMF, its probably quite apt:
Retention is the ultimate test of product-market fit
(Retention: The situationship of SaaS)
As I highlighted from the BI piece on the legal-tech bubble - a lot of what is being booked as recurring revenue may actually be, in part at least, pilot programs, so it's yet to be seen how sticky it is and how much traction they can actually get - which I think comes down to the product market fit question.
I’m not suggesting people don’t love this new cohort of legal AI platforms, or that there aren’t people getting genuine value and efficiency from them. Likewise, I'm also not suggesting that the current vendors aren't going to find PMF, just observing that the industry/market as a whole appears to still be very much in that discovery phase.
(If you are wondering why this is such a short piece, and possibly not that incredible - I have a longer form version of this that I wrote on the weekend, that gets into more market specific stuff, but decided it wasn’t for here. Feel free to hit me up for the longer version, though 😉)
Dentons’ CTO take on the Lexis/Harvey deal
This was another piece I really enjoyed, I can honestly say I’m 100% here for this form of reporting. Loads of good insights, his view from the ground as well as a bunch of actionable things to consider over the short/medium/long term.
It’s great to see him give specific numbers as well as direct implications feature and technology wise (I was going to start quoting my favourite bits, but it was going to be most of it, so just go read it - at the very least read the tables he includes).
YouTube add’s AI Search
A lot of what I write here is basically some insight into the workings of my brain. Usually I see something that in itself isn’t super interesting but then my brain bounces about following random, sometimes long forgot connections and I end up with the muddle of incredible thoughts that you see here.
First of all, the story: Google is adding AI search to YouTube, inline with it’s zero-click AI search (users can search on Google and get the answers they want from Google’s AI search summary, not needing to click further or ever leave Google). The YouTube version is basically the same idea, you can search for something on YouTube and in the results you will also see a similar AI search summary, ideally aiming to extract the key parts of the video (in text form) that answer your query, so you might not even need to watch the video at all!
The summary is intended to extract the information most relevant to your search query, so you may not even have to watch the videos.
Imagine, a video streaming platform trying to help users not watch videos. What a time to be alive.
One thing I find interesting is the cyclical nature of these trends - I commented last year on a report on Gen-Z app trends that they increasingly used TikTok (e.g. video) as their primary search engine, which seemed crazy to me (I usually get irked when I Google something and everything above the fold is youtube videos). But this update will again move people back to text based searching. Much like mobile phone habits went from voice calls (remember back when all anyone cared about was number of “free minutes” on their mobile contract?) → texting only/voice calls are gross → whatsapp voice notes. Or maybe even the renaissance the early noughties indie bands are currently enjoying.
If nothing else, the round & round reference is a chance to link a classic early noughties r’n’b Hi-Tek song..
Beauty maybe isn’t in the eye of the beholder?
I also stumbled upon this article somewhere, about design, beauty and it not being that subjective. The article itself felt a bit like it was mostly a not-so-veiled rant about people not appreciating design is a skill/science, and I’d guess coming from someone who’d experienced the not un-common back-and-forth of I don’t like that colour.
Don’t get me wrong, I agreed with all that was being said, but it felt like the author just wanted to get something off their chest, is all. I did like the section in the middle about how aesthetics re-enforce user trust, very crudely summed up by saying: if it looks crappy, then people probably don’t care that much about this product, so I’m not sure I trust it/want to engage with it.
As I mentioned above, this is another case of my brain wandering through connections made and inevitably thinking about lots of other, very random, things.
First place my brain goes is to the many things I’ve heard/watched about the aesthetic in the Fibonacci sequence, golden ratio and how its used in design and art (I actually once saw a display in Kew gardens based around the Golden Ratio), although a lot of this is slowly being disagreed with by academics, I still think its neat to think about (and still undeniably lurking present in biology/nature). Some of this stuff was popularised by the Dan Brown novel The DaVinci Code, and later the movie, but for me it started with a late nineties Darren Aronofsky movie.
Prior to making pro-AI videos with Google and becoming a multi-award winner with movies like Black Swan and The Wrestler, he debuted with a movie called Pi, a black and white low budget movie about a bunch of people obsessed (for different reasons) with maths and the Golden Ratio. It was fun, and had an awesome ‘90s electronic music soundtrack. He followed that up with Requiem for a dream, which I can comfortably say is the most traumatic thing I have ever watched, and won’t in good faith recommend it to anyone, but to its credit, it birthed what is in my opinion, the greatest soundtrack song of all time, Lux Aeterna.
Other stuff I found interesting
If I still drank beer, I’d probs be tempted by this 28 years later X Tiny Rebel collab. I enjoy a gimick, and like the welsh brewery (although some of their flavours are a bit too nuts for me.
Glastonbury was this weekend, which means I can stream live sets from the comfort of my home (A tonne of people who I can only assume were attending Glastonbury got on my train this morning at Bristol, and when I say people I mean children. I felt old. To their credit they didn’t seem to weathered by the experience). Not too deep into stuff yet but the Prodigy set is my highlight so far. Ezra Collective are always good fun (and one of my recommendations from two years ago). I’ve enjoyed the bits I’ve heard of Doechii’s set, so that one is next one cued up.
Following on from what I was saying the other week about being a hypocrite using AI whilst not agreeing with the companies, this is a very well put together (if not broadly upsetting) writeup of just how bad Big-Tech/Big-AI really is (although upsetting, I’m guessing no one is actually surprised by who takes the top two spots for worst offenders).