Intro: 8 Themes from SXSW 2026

Three Years, Three Characters

• 2023: The emotional reaction – ChatGPT had launched a few months earlier and SXSW lost its mind. Every session was about prompts and productivity. Raw excitement, raw anxiety. Nobody had a framework for what was happening.
• 2024: Framing  – The mood shifted from wow to okay, what do we do with this? An MIT/Stanford article showed a diagram I keep coming back to (see below). The vast blue space is where human and AI capability combine to create entirely new value. The small white rectangle is where most organisations are operating: existing capability, just done faster. Organisations started investing, restructuring, hiring AI leads.
• 2025: Infrastructure – My three takeouts: AI is no longer optional, it’s infrastructural. This era is AI-native, not AI-retrofit. And the next era of design is context-centred, agent-aware, trust-sensitive, and participatory.
• 2026: The emergent AI-class – AI tech has kept accelerating (that’s recursive development for you), but the most interesting talks weren’t about capability. They were about consequence. What AI is doing to our thinking, our organisations, and the structures we’ve built over 150 years.

(Original MIT article ‘the Turing trap)

But, buried in all the things going on, there was a clear divide opening up between people engaging with those consequences and people still treating AI as bolt on.

Too many were pawing at the ‘AI needs humans’, ‘Creativity matters guys!’ and my pet peeve, the ’We’ve seen this before’ types.

This stuff may well turn out to be true, but to me it feels like Hopeium and Copeium.

Note: Cybertrucks are everywhere in Austin.

Eight Themes

This article introduces a series of eight themes, derived from the strongest talks at SXSW 2026. Each will exist as its own article for depth and explanation.

1. The Efficiency Trap. Most organisations are using AI to do the same work better, faster, cheaper, and when AI gives people time back, they use it to do more of that same work. They’re accelerating an obsolete system.
2. Leadership will break in the Agentic Era. When AI executes faster than your organisation can decide, traditional leadership practices and structures won’t be able to keep up.
3. Context Is the Only Moat. AI is commodifying every capability that isn’t proprietary. Your context layer is your differentiator, and context engineering becomes brand gold.
4. Cognitive Offloading vs Cognitive Surrender. AI can augment your thinking or quietly replace it and the neuroscience evidence is harder than most people realise. Head Work Vs Leg Work, the way you choose to work will be formative, and good ol’ hard work is good for you.
5. Trust at Scale Remains Undesigned. Everyone talks about the need for trust in AI systems. I’m seeing companies already spinning it as ‘what they do’. Lots of talk, but little concrete understanding about what it means in practice. To me, this is the near-future for the design industry.
6. From UX to AX. When agents do the work, design’s job shifts from execution to orchestration. Everyone becomes a ‘creative director’ with an army of tireless subordinates. Vision, judgement, taste, pattern recognition, experience all matter — hurrah (yes that’s my em-dash), but DANG, the middle gets squeezed. Think of a conductor: the musicians know their parts, they have the score, but the conductor shapes emphasis, dynamics, timing, feel. That’s where design value lives.
7. Things Will Break. The Question Is What You Build Next. Underneath the strategic and structural themes runs an emotional current: the human experience of being in the middle of disruption. Multiple speakers acknowledged that the next few years will break economic models, creative practices, and professional identities. The response isn’t denial or acceleration, it’s intentional adaptation.
8. The Craft Rewired. The old craft pipeline gave us intuition, taste, and pattern recognition as byproducts. AI is stripping those out. Simultaneously, the floor is dropping on execution and the ceiling is rising on orchestration. What does a designer actually do differently next Monday? This theme gets practical.

Every one of these themes contains a fork.

Which side you end up on is still, for now, a choice.

I’ll be breaking each down over the coming weeks, grounded in specific talks and lived experience. Because these aren’t just SXSW observations.

ELSE has been working on AI projects since 2018, and right now we’re deep inside a couple of genuinely complicated AI transformation programmes.

We’re not commentating from the sidelines.

We’re in the shift.

We’ve been moving ELSE to AI-native for a while now.

Building our own context layer.

Redesigning how we orchestrate human and AI work. Running audits on our own workflows.

Asking ourselves the same questions we ask our clients: are we retrofitting AI into how we already work, or redesigning the work itself?

What does that look like day to day?

People who move between operator mode and architect mode in the same morning.

Who build context layers as naturally as they build wireframes. Who treat curiosity and critical thinking as professional disciplines, because in a world of cognitive surrender, those qualities are what separate the emergent AI-class from everyone else.

We were over 40 people at one point.

That seems crazy to me now.

A small, deliberately assembled team can see further and move faster than headcount ever could, and we take responsibility for the future relevance of our people.

That’s the bet we’re making, because I know which curve I want to be on.

This is the first in a series from ELSE’s attendance at SXSW 2026, if you want the subsequent articles in your inbox, sign-up below. We don’t spam, in fact, we’re far more infrequent that we’d like.