The group selfie with 'The Daniels' after their maximalist "Everything, Everywhere" talk
TL;DR version provided by GPT-4:
SXSW was like jumping into a tech blender set on "future-shock" - a relentless marathon of brain-stretching talks, debates, and showcases from dawn till dusk. Picture this: choosing between five mind-blowing sessions every hour, clocking in enough steps to hike a mountain, all while diving deep into the emerging tech super-cycle.
Amy Webb threw down a gauntlet with a behemoth slide deck, preaching the AI-led tech gospel. AI was the rock star, gatecrashing discussions from film production to fashion, with everyone from Microsoft to OpenAI stirring the pot.
Ray Kurzweil dropped singularity bombs, painting a future where your brain might just be the next cloud server. Amidst the AI hype, cautionary tales and ethical quandaries danced around, reminding us that with great power comes great...oopsies.
Beyond the silicon, there were tales of space architecture that made Mars feel like the next Brooklyn, quantum computing musings that made your head spin, and a revolutionary call to see human hair as the next eco-fab fabric.
And just when you thought it couldn't get any weirder, there were deep dives into internet subcultures and comedy acts poking fun at our digital obsessions.
In a nutshell, SXSW was a wild ride through the looking glass of tomorrow, leaving attendees buzzed on the future, questioning reality, and maybe, just maybe, ready to welcome our new AI overlords.
Things to download and read/watch separately:
- The Future Today Institute annual report
- Amy Webb, on the new tech supercycle
- The OpenAI keynote with product lead of ChatGPT, Peter Deng
- Tom Roach and Grace Kite, doing a very good talk about effectiveness
- Everything Everywhere All At Once, on stories, AI and maximalism
- And Joy Buolamwini, shared her poetry in code doctoral and her piece on Algorithmic Bias “AI Aint I A Woman”
- Joost Van Dreuen on the gaming landscape
- Dylan Mulvaney on a panel on “when beer goes viral: the role of brands fighting hate”
- The CEO of Patreon talking about “The Death of the Follower”
The long version… (Sorry, it's long)
SXSW was a whirlwind of different ideas, presentations and themes. It was great, but pretty overwhelming, every day I covered about 15,000 steps between 8am-10pm, with every timeslot having to choose between 5 equally interesting different talks, and packing more events into a day than I normally do in a month.
This attempts to summarise some of the things I saw and enjoyed, but also with some links to some of the talks and presentations. Some of the big talks are posted on their YouTube channel, so you can get a sense of it all without having to make the trip.
Entering the next “tech super-cycle” with Amy Webb
One of the big reports ‘dropped’ at SXSW is the Future Today Institute annual report.
Free, and coming in at a little under 1000 slides long, it’s not a quick read. But if you manage to find time to read some of it, it’s a pretty robust primer to all of the exciting potential tech trends that are emerging.
CEO Amy Webb, talks compellingly about how all the evidence suggests we’re moving into a ‘tech supercycle’.
Here she was at a smaller session with Fast Company
Previous supercycles have included steam power, or ‘the internet’. Our current one is spearheaded, unsurprisingly by AI, but not limited to it. She talks of three major forces – AI, Biotech and “an ecosystem of connected devices”.
One of the main things linking together AI and XR/Spatial computing is the shift from ‘Language Models’ to ‘Action Models’ – and building working action models requires ever more sensor data.
So, ask not what spatial computing can do for you, but what you can do for spatial computing! This ‘face computer’ is packed full of motion, camera, voice, and eye tracking technology. The next wave of AI development will be underpinned by all of this data, and then in the future it will be powered by it.
This would allow for systems that can help pre-empt or anticipate your needs and your actions, allowing for a truly intuitive agent based computing experience.
Then the ecosystem of hardware becomes less focused on the ‘phone / connected computer’ in your pocket, and will become distributed across a range of devices – lapel pins, smart glasses, ‘face computers’, connected home devices, cameras and microphones. (And devices as yet to be invented, including Brain Computer Interfaces potentially).
All of this takes an imaginative leap – to a world with less swiping and staring, and more voice, gesture, and anticipatory experiences. But a picture that starts to stack up as a potential 5-10 year view of how we might live our hybrid digital lives in the future.
Lastly, biotechnology leans into the potential of AI to make a lot of progress in how we understand biological materials. This not only allows us to solve or improve a lot of biological challenges (disease and aging being the obvious ones). But it also leads to biological computing as a new ‘post-silicon’ era of computing. The human brain is one of the most efficient and powerful supercomputers to ever exist, and it’s a potential strand of future computing solutions.
The future of AI is sprawling, hyped, and happening (whether we like it or not)
Almost every talk that wasn’t about AI, was still about AI. AI was talked about as the great enabler. Suddenly new frontiers of XR, Creativity, Enterprise and Industry, Robotics, Film Production, Journalism, Fashion, Music, Medicine and more, were possible.
Several of the big talks were from key folks working in the AI space. There was representation from Microsoft, Deepmind, Amazon, HTC, Intel, and, of course, OpenAI.
The OpenAI keynote with product lead of ChatGPT, Peter Deng, was (unsurprisingly) full and had people queueing out the door. It was also a very compelling and well-rounded discussion on the current opportunity and vision of AI and product development roadmap. It was a tad utopianist and he somewhat ducks the difficult questions about creator rights, ethics and responsibility. (Although nothing as bad as the interview with OpenAI CTO struggling to explain what Sora is and isn’t trained on.)
Clashing with it, was Tom Roach and Grace Kite, doing a very good talk about effectiveness, and how start-up brands need to shift and learn to ‘scale up’ as transition off the drug of performance media quickly. In this talk, it looked at the science behind performance and brand, for different stages of a brands growth. It also was interesting how it explored the journey Amazon went on to embracing advertising – from an early quote from Jeff Bezos “advertising is the price you pay for having an unremarkable product or service” to becoming one of the biggest (and best) advertisers, but also building one the biggest advertising networks in the world.
The Godfather of AI Singularity was there
Ray Kurzweil seemed both overwhelmed and bemused by the turnout for his talk. Unlike everyone else, jumping into the AI fray, he’s been talking about his vision and roadmap for AI since the turn of the century. In the past, he was often dismissed as a fantasist. The Singularity claims that when we get machines that have exceeded our intelligence, when we can map our brains and replicate them (memories and all), and apply recursive improvement through interconnecting and sharing our knowledge, we’ll reach a mass consciousness event like a black hole, whose metaphysical gravity will be such that humanity will never be the same again. (Ok, when I put it like that, it does still sound quite sci-fi…) But Ray predicted we’d reach AGI or Superintelligence around 2029, and singularity by 2045. Recent developments put us potentially quite on track for that, and his updated new book “The Singularity Is Nearer”, launching soon, will say as much.
AI, Applied
There were also some interesting academic progress being shared. Things like weather prediction sat between Quantum Computing and advanced AI modelling – a useful reminder that while LLMs are powerful, they’re also only a small part of the AI puzzle and potential. Corpus Christi used AI to save over 10,000 turtles when it predicted cold weather fronts that would cause them to rise to the surface and get harmed.
While, Zach Seward from the New York Times, talked about his experiences of AI in journalism – the good, the bad and the ugly. His presentation here, contains the examples, which were an impressive array of applications of custom AI to raise the abilities of investigative journalism. And if the tools are good enough for investigative journalism, they should work for strategy, insight and creativity.
One of the most interesting, was the Wall Street Journal training a model to identify lead cabling around schools in New Jersey, from trawling Google Streetview images, allowing them to understand the scale of the problem, get a breaking news story, but also offer a solution.
Recommended by LinkedIn
Some of the best talks were cautious or critical of AI. The Daniel’s, director duo of
Everything Everywhere All At Once gave a hugely eloquent and engaging tour-de-force on maximalism and the human condition. Here are some quotes that captured elements of it, but it’s also available online and a must-watch for fans of the film, or humanity:
- Chase the contradictions - that is where the universe is showing you its seams. It’s inviting you to come there to play
- The world is maximalist. We wanted to talk about the whole world but it’s impossible. That’s why we needed to make a multiverse film that can tell this story
- Stories can be a compression algorithm - and stories do this all the time. Thousands of stories all connected, inter-dependent.
- The system we have now are just fossilised stories of the past, and as those stories start to crumble, we have a chance to undo what the stories have done to us
- AI is incredible. It may be the tool we need to cure cancer, give climate solutions. But, what stories will we tell ourselves to adopt this? What will this system do when we add AI to the mix?
And Joy Buolamwini, shared her poetry in code doctoral thesis on the algorithmic bias in facial recognition technology and the risks of bias across the whole system. This culminates powerfully in her piece “AI Aint I A Woman”. Her MIT thesis ‘Gender Shades’ proved that FRT algorithms are fundamentally biased and causing very real harm. This action has led the platforms to improve their technology, and also to stop using it for policing (which has led to several false arrests and imprisonments). The work setting up the Algorithmic Justice League is fascinating, and the concept of ‘coded gaze’ and being ‘excoded’ are two powerful concepts. (‘Coded Gaze’ is the machine equivalent of ‘male gaze’, while being ‘excoded’ is when an algorithm excludes or denies your identity).
She also had a point that our trust in algorithms is often too strong, and truly automating things will be harder than we think. One really great (terrifying) example she shared was of the National Institute of Eating Disorders, which fired all their call-centre staff and replaced with a chatbot, which they had to roll-back only 2 weeks later once they realised the chatbot was giving ‘eating disorder enabling advice’.
XR and Spatial were the biggest other thing since ‘the metaverse’
The hosts of ‘This Week in XR’ podcast, had a panel with the founder of Magic Leap, Rony Abovitz, and Joanna Popper, previously ‘chief metaverse officer’, where they penned a “Dear Tim” letter to Tim Cook about their feedback (with love) on the Apple Vision Pro.
The general take was that it’s ‘an expensive and impressive work-in-progress’. It’s too heavy, too front-weighted, too expensive. But, that for certain types of immersive entertainment, it could be pretty breathtaking, and that we’re yet to really
Rony was also pretty damning that it’s essentially a VR experience and that the weight and the experience of the passthrough technology, is the wrong avenue for developing this. And that the more promising angle is something like XREAL which are lightweight glasses with a screen overlay. (See-through vs Pass-through – pass-through is potentially a bit of an XR cul-de-sac, where you’re building an incredible screen that can recreate the world, rather than act as an overlay on the actual world).
Ultimately, the components of spatial computing need to entail a camera, a microphone, a connection to a computer and/or the cloud. So this could end up being a combination of watch, lapel pin, glasses, and yes, probably a mobile phone in the pocket. And that also becomes an AI powered software / experience challenge – linking together all of these devices into a spatial experience will be tricky for anyone, although a potential huge strength for a well curated walled garden like Apple.
New frontiers of solutions
There were great talks (and some less great panels) about the future of computing, robotics, XR and spatial, space exploration. I won’t cover them all, but a couple of interesting highlights.
BCI (Brain Computer Interfaces) for both medical and augmentation – the next 10 years will see some breakthrough in how we can understand the brain, train computers to understand the brain, and connecting them together. The nearside area with this is through medial application – ranging from understanding reactions to drugs, to providing non-chemical brain treatments, to helping those with diseases like ALS to connect with the world through robotics and virtual world simulations.
Quantum computing and non-typical computing approaches – quantum computing is complicated. (Duh). But, there are multiple companies getting close to a useful working model. Quantum has a strong theoretical potential to be very good at complex computations, and probabilistic computing (which is things like modelling the weather).
Space architecture was a talk I stumbled upon when the main talk I wanted to get into was full. So instead there was a talk inspired by Buckminster Fuller: Space Architecture for Spaceship Earth.
"There are no passengers on spaceship earth - we are all crew" - Marshall McCluhan
This looked at some of the systems thinking and technical and design solutions needed for us to build permanent settlements on the Moon, Mars (and beyond!) – an interesting area that is less speculative fiction than you might think. However, the most interesting twist, was how the constraints of space architecture, could also be inspiring or instructional for earth architecture – and that breakthroughts in materials and processes could lead to less wasteful, more sustainable, closed loop production and construction methods for Earth.
Digital twin production for Dune 2 also looked at the way that virtual world simulations could help with real world production. Denis Villeneuve shoots for real as much as possible. So while the ornithopters and space landers were VFX, the shoots in the desert, and the explosions were real. However, shooting accurately in the desert brings some challenges – one in particular was shadows and shade. The production team visualised all of the different parts of the locations, so they could model when the sun would be and how they could work with the shadows.
Certain scenes blended 9 different locations and times of day in one scene
They could also put Denis into VR to preview the locations, the scenes and the prop design – for example seeing what the finished ornithopter interior would look like. Which felt like a pretty cool use of Unreal models for real-world production.
There were hundreds of other tracks, in music, creativity, gaming, psychedelics and culture
Ranging from talks on Gen Z attitudes to sex scenes in movies, to the music industry and fandom, to medical research with the Chan Zuckerberg Institute. Dylan Mulvaney a panel on “when beer goes viral: the role of brands fighting hate” and the CEO of Patreon talking about “The Death of the Follower” (as a concept, rather than literally) are both worth a watch.
I didn’t make it to any psychedelics talks (of which there were a lot), but I did make it to some gaming talks. Joost Van Dreunen did a great talk about the state of the gaming industry, and helpfully also shared his charts.
And there was even a talk about Human Hair as a new and overlooked material. In one of my favourite talks, we heard about how much hair is thrown away at salons around the world (2.2b kgs), and how this can easily be processed into a hugely versatile, robust, water resistant, anti-bacterial fabric.
Hair, after all, is just human wool. Why do we love our hair when it’s attached on our heads and think it disgusting when it’s cut off? Why do we think of wool as ‘clean’ and hair as ‘dirty’?
If it had been a Dragon’s Den pitch, at the end I’d have shouted “I’m In”. Their plan is to scale up their factory to produce 500 million kg of hair over the next 10 years, both saving that waste to landfill, but also reducing the need to produce new material elsewhere. It wont replace cheap polyester, but it brings a new part of the circular economy into play.
It wasn’t all talks… The music, film and comedy all had something to bring to the table
While I wasn’t there for the film festival, I did go to the screening of “The Anti-Social Network” coming to Netflix in April. This was a terrifying tour-de-force of internet culture, specifically focusing on 4chan, its origins, and its impact, ultimately leading and contributing to Qanon and the storming of the Capitol Building on 6th Jan.
If you want a fascinating deep dive on unintended consequences, the dangers of anonymous online communities and the way that misinformation spreads, it’s pretty excellent. Documentaries are kind of like extremely well put together keynotes that can bring something to life in a deep new way.
And the comedy also brought some incisive storytelling. One comic railed against the dangers of nerds ruling the world. As we look at the algorithmically powered attention machine, capturing our data and our eyeballs, nudging us to scroll more, want more, and worry more. Maybe the jocks were right?
Alongside this, it was also a space to premiere and launch big new shows for the year – with big activations for Netflix 3 Body Problem using special build projection hologram (apparently by Holotronica – a South-West England based company – so truly South By Southwest…!)
Fallout, a TV series launching on Prime based on the game, also built an immersive experience on the outskirts of town (and had branded busses ferrying people to and fro).
So – to summarise, there was a pretty consistent theme around a new “tech super-cycle” emerging, underpinned by AI, but also looping in a world of spatial devices and technologies. There’s undeniably some over-hype around AI, and some concerns around its impact on the world. But there was a definite energy of experimentation and new possibilities emerging.
