Small Is the Next Big Thing

I first attended South By Southwest 15 years ago.

Each year there has been one meta theme. Things like security, gender politics, robots, rockets, medicine. All based around a central technology: The Internet.

Then three years ago Sam Altman turned up and talked about what he’d launched three months earlier.

The Chat GPT was out of the bag.

The buzz back then was all about Artificial General Intelligence; the nightmare-ish dream of machines matching human intelligence. No one knew where things were headed but everyone knew the world had changed forever.

The following year everyone was sprinting to be the next big thing in Generative AI. An entire industry had sprung up to offer endless ‘solutions’ of greater or lesser use.

Now, two years after AI hit us, the hysteria has died down and attitudes are maturing.

AI was nowhere near the sole subject of every session in Austin this year, but they all had one thing in common.

Or rather, two letters.

Niall Firth of MIT talked about the biggest existential subject facing us and one the tech industry needs to get a grip on: sustainability.

He explained that the launch of Deep Seek has proved you don’t need masses of environmentally damaging computing power for a LLM.

Small is the next big thing.

Because you don’t need the entire internet for most questions the future lies less in expensive and power hungry LLMs but in Small Language Models (SLMs), trained and fine-tuned on a company’s own data, they will be more task-oriented, more efficient, less expensive and better for the environment.

He also talked about Gen AI Search soon making 'Google seem like silent movies'.

Sadly, it’s not great news for journalism. This new era of ‘Zero click’ with AI serving up news summaries means people won’t need to visit newspaper sites.

Christian Amon, CEO of Qualcomm, one of the biggest players in semiconductors, talked about the world’s profound move from humans understanding computer language, to computers understanding human language:

This means our voice is becoming the new UI, ultimately replacing the keyboard. What Alexa and Siri have been trying to do will become widespread. And will work

This way of communicating with our computers alongside AI Agents will ultimately replace apps: your AI Agent will suggest actions that apps currently do.

They’ve talked about wearables for years at SXSW but this is the first year that things like Meta glasses may take off. That is, if people can cope with having to wear them constantly, which remains a massive ‘if’.

It was interesting how rarely the tech giants Google, X and Meta were referenced. One exception was when the fearlessly entertaining and erudite Professor Scott Galloway summed up what 1700 people in the room seemed to agree on: “Mark Zuckerberg is responsible for everything that’s bad in America”. This received the biggest round of applause of the day.

Among the many insights Prof G threw at the audience was the fundamental human truth that people want fewer choices. And that the algorithms of brands like TikTok actively take choice away, which is what makes them so popular.

“Google gives you a massive selection. AI gives you what is right for you.”

His three big media predictions for 2025?

YouTube will be the hero streaming platform of 2025, turbocharged presumably by AI’s increased accuracy in serving us appropriate stuff rather than any old crud.

Podcasts will continue their huge growth, being one of the few ways toAI

The value of Reddit will increase as it becomes the arbiter of the best podcasts, helping narrow people’s options.

Quantitative Futurist, Amy Webb, talked about the effect of AI on biology.

“Any company that makes any physical product will be affected by this.”

We’re going to see products engineered with ‘meta-materials’ that break the normal rules of physics. Bricks that are flexible so buildings don’t break in earthquakes. Lab grown rhino skin draped around cars to withstand crashes better than any material we have currently.

The future is LI: ‘Living Intelligence’.

We’ll see programmable organic digital networks, human brains and silicon chips with living neurons, brain parts in living computers.

We should also expect wearables for our cells. ‘Spermbots’ for example. A sperm cell wrapped in a coil so a magnet can force it to go directly towards the egg.

Ben Lamm, CEO of Colossal Biosciences made the horrifying announcement that that we are on course to lose 50% of the world’s biodiversity by 2060.

His company’s goal: to make extinction a thing of the past.

AI has transformed the work they do – editing genomes to add mammoth DNA into mice, growing hair in Petri dishes, creating artificial wombs and, he hopes, reintroducing the dodo in Mauritius.

Within the next 20 years, he says, we’ll eradicate 90% of diseases, have a cancer vaccine and 'achieve longevity escape velocity' – living forever.

But before we got too excited, he added that AI plus microbiology is potentially the most dangerous technology the world has ever seen, leading to a new arms race.
This year, we saw two great female tech CEOs shining above the phalanx of tech-bros.

Jay Graber, is CEO of Bluesky which she describes as a “customisable social network that allows people to keep their own data and identity”.

Think Twitter, but not run by a maniac, where the algorithm doesn’t dictate what appears on your feed, you do. Allowing you to choose your own moderation tools, avoiding the misinformation and hate posts.

She declared the app 'billionaire proof'. Its open source code and business model means it wouldn’t be attractive to a billionaire.

Meredith Whittaker, the impressive CEO of Signal, sprinkled a hefty dose of paranoia, explaining that “the economic engine of tech is surveillance and the only way to protect people’s data is not to collect it”.

All user data on the servers of all the big tech companies, she said, could be used by bad governments. Her company, having refused to put in a back door to its app, makes it the ONLY platform on which you can send messages confident that they won’t end up being read by people or governments wanting to do you harm.

And she outlined the security issues over Agentic AI. We’ll be unwittingly giving the tech companies access to our browsers and our credit cards, calendars, messages – letting bots scavenge every aspect of our lives.

Fifteen years ago, when I first came to SXSW, Google announced its foray into autonomous cars. The audience was intrigued, excited but more than a little sceptical.

In the preceding years we got a few updates of the millions of miles the cars were driving around the town Google built specifically to test its, eventually renamed, Waymo vehicles. It still seemed like a pipe dream.

How poetic then when, during this year’s South By, Waymo launched on Uber.

Waymo robotaxis are now in a handful of cities clocking up 200,000 paid trips a week.

Jointers loved their rides around Austin and didn’t have any safety concerns.

It feels like the future has arrived and is parked outside waiting for us to jump in.