Gen AI's Act Two starts here.
Anthropic and Adobe offer contrasting visions. Which should you choose?
This week Anthropic and Adobe gave us a glimpse of the future(s) of Gen AI.
Note the all-important ‘s’ – futures.
In 2017, long before the current AI boom, I co-wrote a piece about how the future of customer experience would be defined by your answer to the question: are you saving attention, or seizing it?
Now, Anthropic’s computer use model and Adobe’s Project Concept show how a similar divide is opening up around where and how you’ll use generative AI:
Are you deploying AI for convenience – focused on automation, productivity, and efficiency?
Or are you using AI to augment and amplify human creativity – crafting richer or even totally new experiences?
Let’s dive into these two breakthrough products in more detail.
What do they tell us about generative AI’s Act Two – and, more importantly, what they’ll mean for you in 2025?
A note from our sponsor (me!). You’re probably deep in 2025 strategic planning, sales kick offs, and customer conferences. 🚨 I’d love to help.🚨
Typically I inspire clients with:
Your Future Normal: a trend keynote, bringing you relevant yet non-obvious insights from outside your industry.
VisuAIse Futures: a new interactive, ‘multiplayer’ AI-powered experience to unleash the creativity of your team. Watch the 2-min highlight video.
Read more about these at the bottom of this email, if you’re interested.
In the last few weeks I’ve had gigs in London, Bournemouth, Hamburg, Berlin, New York, Las Vegas and Greece. Still to come are Rome, Dubai, London and Berkshire :)
I have a few slots left before the end of the year, and am taking bookings for 2025.
If you’d like to discuss bringing me to your next event, please do reach out to Renee Strom on renee@ideapress-speakers.com.
Anthropic’s computer use & AI agents
Even if you’ve tired of the AI hype machine, you’ll probably be aware that agents are AI’s “Next Big Thing”. In fact, I dedicated the last edition of this newsletter to “The Weird World of Working With AI Agents”.
This week, Anthropic released its computer use model which can, “follow a user’s commands to move a cursor around their computer’s screen, click on relevant locations, and input information via a virtual keyboard, emulating the way people interact with their own computer” (my emphasis).
In short – now it can get sh*t done.
Just watch the video below. Claude plans where to watch the sunrise, researches the drive time, and then adds it as an event to the person’s calendar:
(Or if you must, watch the business-focused use demo video where Claude fills out a vendor request form).
There are countless pieces dissecting these new capabilities (two good ones are Ethan Mollick’s “When you give Claude a mouse” and Simon Williamson’s “Initial explorations”). I’m certainly not going to (or am I remotely qualified to) get too into the technical details here.
The tl;dr, we’re very early. The consensus is:
It’s still very much a developer feature.
Real-world reliability is a challenge.
Security is a real issue.
BUT. Watching the demo videos it’s clear that the Next Big Thing is quickly becoming the Now Big Thing.
Anthropic has been a pretty level-headed player in the AI game, so it is especially powerful to read, “we were surprised how rapidly Claude generalized … [it has a] remarkable ability to turn a user’s written prompt into a sequence of logical steps and then take actions on the computer.”
Despite this, my main reason for writing about Claude’s computer use is what it can’t do. I don’t mean practically, but conceptually. And most important – what it means for your AI strategy1.
I touched on this in The Weird World of Working With AI Agents, but Claude’s computer use model highlights a key truth about working with AI agents:
Autonomous AI agents will be the future normal for ‘known’ tasks – those where you can define what a successful result looks like.
This leads us to the paradox of AI agents – the more capable and autonomous they become, the greater the need to define what success looks like. Otherwise, as Ethan Mollick notes, they can quickly end up performing excessive or even infinite iterations, or inadvertently causing harmful outcomes.
And this is just the start. A few further provocations to consider:
Tasks done on computers, will be done by computers. Coding. Bureaucracy. Reporting. Testing. Executing transactions. It’s now clear that any computer-based task, will (eventually) be done by an AI agent. The big question is what volume of currently ‘offline’ tasks can be digitised? We digitised money. We digitised music. We’re about to digitise mobility, with self-driving cars. What’s next?
Prompt Engineers are dead! Long live Agentic Auditors! Yay, another consultant goldmine! ;) More seriously – if we’re no longer coaxing LLMs via iterative discussions then it will be critical to be able to check if agents are secure and on the right track. However, like prompt engineering, over time this will become largely seamless and invisible (normal) to the mainstream end user.
Efficiency > Ethics? Or vice versa? ‘Success’ is rarely one dimensional. Which means ever more powerful AI agents will lead to ever more thorny ethical questions. A simple example from the video above – should the AI pick the fastest travel method and route, or the most scenic? Or the healthiest? Now imagine a procurement agent – what happens when the cheapest option is lower quality? Or less sustainable? Will we need to keep humans in the loop to make these calls? Especially as our preferences aren’t static.
As we edge into a world where AI agents increasingly automate our ‘known’ tasks, we’re going to have to ask ourselves some fascinating questions. Indeed, all of two weeks ago I published 10 further questions about our agentic future which remain relevant. For now ;)
Adobe’s Project Concept & augmented creativity
Automation is just one half of the AI puzzle. What about creative tasks, with unknown outcomes?
Back in late 2022, as ChatGPT burst onto the market, we wrote a chapter on ‘Augmented Creativity’ in our book, The Future Normal. Indeed since then I’ve written many times about how ‘Crowd-Powered Creativity’ will be one of the most profound and disruptive implications of generative AI. I’ve even built my own multiplayer AI creativity tool to bring audiences’ insights into my keynotes.
Which is why I’m completely obsessed with Adobe’s Project Concept – perhaps the purest manifestation of how AI will augment, amplify and accelerate human creativity I’ve seen so far.
In Adobe’s words, it will “help you rapidly explore potential artistic directions, mix images together, transform regions of an asset, and remix styles, backgrounds, and other ingredient assets. Project Concept uses AI and collaborative tools to assist with both divergent thinking, illuminating the far-reaching possibilities of any creative project, and convergent thinking to bring ideas together, apply controls, and ultimately select the final direction to go with.”
The AI-powered future of creativity
Generative AI has polarised the creative community. I take a slightly more nuanced view than Fiverr – it’s not that ‘Nobody Cares’. They absolutely do. But people won’t care in the future, especially if copyright issues can be resolved.
In the future normal, generative AI will be to creatives what spreadsheets are to accountants. Done right, it’s simply too powerful a tool to resist:
Augmenting > Automation. Theoretically, AI agents could generate, test & iteratively optimise even ‘unknown’ creative outputs without human involvement. But that ignores the very real costs involved (whether that’s 'just' time, and/or real world production). When faced with infinite options, human taste and expertise will be shortcuts to success.
Innovation at the intersection. Remix culture. Mashups. Fusion. Whatever you call it, innovation happens when we combine things in novel ways – and Project Concept unlocks completely new ways to bring multiple ideas together.
More inputs = better outputs. Project Concept also makes it trivially easy and fast to explore new intersections. Speed matters. Scott Belsky, Adobe’s Chief Project Officer, has written compellingly about how AI gives ‘humans more cycles for exploration, ultimately yielding better solutions’.
Diversity, or democracy? The multiplayer dimension of Project Concept fascinates me. We’ve seen how making it easier to ‘be creative’ has profound implications – not just on outcomes, but also on innovation cultures. Optimistically, this means more diverse perspectives. Realistically, accommodating these can be challenging (or even undesirable).
Beyond visuals. Project Concept is a visual tool, but what if these ideas were translated into other creative mediums? Google Lab’s TextFX collaboration with the rapper Lupe Fiasco, is one attempt. OpenAI’s Canvas and Cove’s AI-powered cards and boards also offer new ways to explore text-based ideas.
Creativity → Brand → Identity.
Generative AI risks ‘flattening’ creativity. Perhaps you feel this in your feed already. However I believe this will create a backlash – having a strong, ‘spiky’ point of view will become increasingly valuable. Amid abundance, we seek scarcity. We seek belonging.
Where Next? Serenity in the Age of AI…
Like every good report, I’ve laid out two competing visions of the future: Anthropic wants AI to do your work. Adobe wants AI to amplify your creativity. Both are right.
Typically, I’d end with a rousing call to action here – encouraging you to consider the nature of the task you’re focused on, as well as your organisation’s purpose and positioning as you choose a path (of course, with the unsaid message: let’s talk if you need help deciding ;)
However while writing this piece with my AI copilots – ChatGPT, Claude and Perplexity – I found myself thinking of the Serenity Prayer.2 So, with apologies to Reinhold Niebuhr – I want to end with a new version, updated for the Age of AI. Enjoy!
Can I inspire your team to seize the future?
This year I’ve delivered 25+ sessions, both live and virtually – from Brazil to Saudi Arabia, Slovenia to Shoreditch.
My regular trend & innovation keynotes bring fresh cross-industry, people-first perspectives to your audience.
I’m also hugely excited about the reactions to my newest offering – VisuAIse Futures.
It’s an interactive, ‘multiplayer’ creative experience that will leave your audience thinking differently about AI:
“It was so refreshing to hear how AI can be used to power human imagination, rather than replace it. And then it was even better to actually experience it”
“Fantastic session! Hugely insightful and fun, too!”
“Brilliant. The feeling in the room was positively intense whilst the images were coming through!
Feel the optimistic vibes it will bring to your event in the 2-minute video below.
If you’d like to discuss bringing me to your next meeting or event then please do reach out directly to Renee Strom or check out my speaking site.
Thanks for reading,
Henry
Will 2025 be the year it becomes meaningless to even think of ‘our AI strategy’? It will just be ‘our strategy’.
Interestingly while it was helpful to explore some ideas with ChatGPT, I ended up going with my own words. Taste matters ;)