Two and a half years ago, no-one was talking about synthetic users, AI-generated UGC or AI agents. What will we be discussing in 2027, that is not even on the horizon today?

This question struck me today, so I made a quick 'AI Futures' experiment, getting Navar*, ChatGPT and Google Gemini to go head-to-head in making some predictions. Below are my four favourite ideas from all of the models…

Right now, AI agents are like an early-stage gold rush. Everyone knows there’s something huge here, but no one’s quite sure how to extract the real value.

Make no mistake – agents will take off. And when they do, they’ll create a whole new problem: the need for 'consumer agents'. Why? Because the flood of AI-driven communication across every platform will be so overwhelming that we’ll need AI just to manage it.

But here’s the catch: today’s AI agents aren’t yet ready for the masses…

Maybe in the generative AI era, the furious pace of development and off-boarding of employees will be accompanied by a renewed interest in long-term thinking and societal responsibility.

Panasonic might be about to give us a demonstration of the perfect balance.

At CES this month, Panasonic launched a programme to "promote the transition to an AI-based business model". But better yet, and perhaps unexpectedly, this is part of its 250 year plan…

The truth is that generative AI should be ‘for everyone’ - especially since one of the biggest shifts is the way it allows you to interact fully in natural language - no coding skills required.

But while there's an ocean of articles, posts and videos 'explaining' AI - much of the terminology used is blocking people from realising how simple it is to get started and get meaningful results.

Let's take one example to illustrate the point…

HeyGen just announced 500+ community-generated AI avatars, which is really exciting progress. But it got me thinking about the future of content creation and fair compensation.

In medieval times, serfs worked the land and paid a 'tithe' to their lords. This could easily be translated to the burgeoning AI content ecosystem...

Deep Seek, the new Chinese LLM, is impressive. But don't jump ship from your preferred AI tools and models just yet - here's why.

Yes, Deep Seek is impressive, especially considering it cost far less than conventional models (DeepSeek V3 was trained in around two months at a cost of US$5.58 million compared to an estimated $41 million to $78 million for ChatGPT 4.)

But in this experiment while it gave a functional response, it was not fully usable, nor as nuanced as ChatGPT 4o or o1….

If your business performance relies on SEO or SEM, prepare yourself for shock and disruption.

One day very soon the trickle of people using genAI search tools will become a flood.

Which means your company will need to target LLMs - in other words ‘market to machines’ - rather than the infinitely more predictable search engine algorithms we’ve been used to dealing with.

And just like marketing to human beings, there is a lot of imprecision and unknowns in the mix when it comes to LLMs.

Germany is f***ed. VW closing three plants, Porsche profits down 40%, Mercedes in trouble, budget deficit for 2025 hitting €13.5bn (due to lower tax revenues).

To stop the tail spin, investment is needed (to fuel innovation in the industries of the future) - but the current coalition is dysfunctional and will likely not agree on sufficient measures…

This week’s news about potential VW factory closures sent shockwaves far and wide. How will Germany fare if its stake in the automotive industry declines, and the next wave of innovation is one in which it has no stake?

We could be witnessing the dawn of a 'new wave' in hardware/software ecosystems. After cars and smartphones reshaped our world, AI and AI humanoid robots could dwarf both industries combined...

Now you can leverage multiple LLMs, for free, to help you get the best genAI deliverable while you experiment and choose your preferred infrastructure.

HuggingChat gives you a chat-based interface AND the ability to select from a set of different LLMs. You can run at least a few queries for free (I didn't hit a limit yet).

I wanted to see how they compared, and at the same time understand how well LLMs might help to implement a content strategy for the upcoming redesign of our 55BirchStreet website. So I chose a hypothetical example of a blog post about genAI and sustainability as a concrete example…

Even in a self-proclaimed 'AI business', understanding how to leverage genAI is not always widely distributed across the organisation.

Here's a LinkedIn job ad for a Group Product Manager at DeepL, a company I use regularly, value highly and respect greatly.

They have a great mission, "to break down language barriers for businesses and individuals around the world", and position themselves as "the world’s most advanced Language AI".

Yet when it comes to recruiting for a relatively key role in their product team, there is no mention of genAI in the job description, neither under Responsibilities nor Experience…

"I never use Google search for anything any more." 18 year old at a gallery during Berlin Art Week yesterday.

Basically, search is broken.He may have been exaggerating slightly, but I honestly get his point.I wanted to book a Berlin-Southend flight to visit family. I know the direct flights were pulled a few years ago, and wondered if they had been reinstated by now.

A 'quick' Google. 61 results on the first page alone. The actual answer is in there, but it's so buried as to be almost invisible…

The meaning of 'team' in 2024 has radically changed.

Collaboration, motivation and delegation - they are all still highly critical skills - but the context in which they are applied has expanded beyond human colleagues, to include a selection genAI 'collaborators'.

Is this really happening at scale today? It's tempting to think it's a case of 'the future is here just not evenly distributed', but that would be underestimating the growing waves already crashing on the shore.

Let's take two examples…

Last week's unveil of the Figure 01 was jaw-dropping. But the implications of industrialising this technology are even more breath-taking.

The initial demo already made it feel like we're living in the future. A humanoid robot that is dexterous, intelligent, can see, hear, converse, follow instructions and reason.

But think about it. With the capabilities that Figure has already developed, they've built their own flywheel.One of the first tasks at scale that the 01 can be deployed to do is to build more 01s…

In the vibrant yet cutthroat advertising landscape of San Francisco, two agencies, Quantum Creative and Narrative Edge, embarked on markedly different paths with the advent of generative AI, setting a dramatic stage reminiscent of a David Fincher narrative—complete with tension, intrigue, and an enlightening twist….

Most people had never used a mobile phone in 1984. And most of them wouldn't have wanted to.

Mobiles of that era had huge battery packs that you had to lug around, and were prohibitively expensive for the majority of consumers - pretty much strictly 'business only'. Basically they were for car-bound, high-powered executives and sales people, for whom being in constant contact made the expense and inconvenience a price worth paying.

If you'd stopped people in the street and asked if they wanted, needed, or were planning to get one - you can imagine the likely response.Fast forward to 2024. Now generative AI is in the same place as those early mobiles…

There seems to be a lot of panic on social media about the impact of generative AI on the creative industry. This was triggered by Sam Altman being quoted as saying that generative AI will do "95% of what marketers use agencies, strategists, and creative professionals for today."

Immediately and understandably, many people’s first reaction was that this sounds like a race to the bottom, where the primary currency is 'fast and cheap' as agencies struggle to cut costs and survive.

But it doesn't have to be this way. There are some really interesting alternative strategic choices ahead…

I’ve been wanting to try this for a while, ever since I got my HeyGen account activated last year, and am very excited to finally have time to conduct and share my experiment.

I was originally planning to try it out with a custom avatar based on my own image, but I decided to see how far I could get using free tools only. In some ways, that proved to be an even more fascinating journey…

The signs are that AI-generated content of all types will explode in 2024. The chart below is just an indication, but Amazon last year already had to limit the amount of self-published books an author could release - to 3 per day.

Yes - you heard right, 3 books per day - which means some people were releasing a lot more than that until the limit hit.

Given that AI struggles to detect AI-generated content, we will need a different solution to ensure we're not drowning in fake news, reviews and machine-generated books, articles, music, youtube content and podcasts…

I wonder if there will be any more 'why did it take so long?' moments in 2024. And for once I'm not talking specifically about generative AI, rather the shift in perspective needed to see where huge value can be created by something relatively simple.

Wheels on suitcases. First patented in 1945, the modern version patented in 1972. It took 15 years from then until it really started to take off, maybe helped by the addition of a retractable telescopic handle. (And maybe delayed by lingering male attachments to the demonstration of strength and masculinity.)

The 'why did it take so long?' feeling hit me again recently when I used a vacuum cleaner with an LED light built in for the first time. The light makes the dust stand out like nothing before! You don't miss any, and you get much more 'task satisfaction' having seen those easily-overlooked dust bunnies whisked away.

It's not as if people weren't trying to innovate…