Davos AI Insights Pt.1: The Employment Shock, Creativity Crisis and Software Revolution
We parse the key insights from five of the biggest minds in AI and consider how they impact our work
1/ Big employment changes are coming - the consensus is clear
The most plausible outcome to this eternal debate atm is Dario Amodei of Anthropic’s scenario of high GDP growth + low employment. (US data on GDP and payrolls already shows this pattern emerging.)
Amodei - whose Claude Code is already multiplying the productivity of the developer community - sees the following:
Different to previous technology paradigms because AI does “anything a human can do remotely”
GDP growth 20%+ but unemployment rises significantly
Government intervention necessary for distribution
Near-term: entry-level jobs disappear (1-5 years)
“My view is the signature of this technology is it’s going to take us to a world where we have very high GDP growth and potentially also very high unemployment and inequality... We’ve never had a technology that’s this disruptive. So, the idea that we could have five or 10% GDP growth, but also 10% unemployment, it’s not logically inconsistent at all.”
Nvidia’s Jensen Huang sees task automation easing bottlenecks, using AI-assisted nursing as an example (though hospitals aren’t exactly sitting empty):
“[B]ecause you could now see more patients and we’re no longer bottlenecked by the number of nurses, more patients could get into the hospital sooner. As a result hospitals do better, they hire more nurses.”
Elsewhere, Huang talked up an infrastructure boom - primarily for AI infrastructure itself, though this could extend to, say, hospitals longer-term:
“This is the largest infrastructure buildout in human history. That’s going to create a lot of jobs. And it’s wonderful that the jobs are related to tradecraft. We’re going to have plumbers and electricians and construction and steel workers...”
The irony being, AI creates massive demand for physical tradecraft while threatening knowledge work. (News broke just this week that Claude is releasing a tool that performs the tasks of a financial analyst, just in time for Wall St to put this to the test in earnings season).
There was consensus from Amodei and Google’s Demis Hassabis that entry-level roles will be hit hardest, first:
Amodei Warned on Entry-Level Roles: “Half of entry level jobs in maybe 1 to 5 years, just gone.”
And Hassabis Agrees: “I think it’s going to be the more entry level things that will go first.”
The pattern: Everyone sees transformation. Most see opportunity. But serious voices acknowledge displacement risk that historical analogies won’t solve.
Our takeaway:
This advantages nimble operators with experienced senior staff. Here, for example, we’re already building simple Claude Cowork implementations for clients to implement internally on their workflows - such as reformatting social content across platforms, that big agencies once charged junior rates to deliver.
AI is taking fat out of bloated budgets. But the problem is, if agencies can’t monetise junior roles, they won’t employ them. So, where will tomorrow’s consultants come from?
2/ No one wants to talk about AI ‘creativity’
Among five major voices - Amodei, Hassabis, Nadella, Musk, Huang - creativity barely registered. No mention of AI slop (other than Amodei’s implication that this is one of the reasons he’s happy to be B2B-focussed, rather than hustle with OpenAI in B2C).
Perhaps the best place to start here is with comments made by the historian, Yuval Noah Harari, in his Davos address. He elaborated a ‘words vs. flesh’ dichotomy with the insight that “AI can describe love perfectly but has zero evidence of feeling it.” Here are his key points:
Words have a new master: “Until a few years ago, nothing on earth could use words. Only humans... Now there is something that is able or soon will be able to use words better than us.”
Power over words is real power: “If laws are made of words, then AI will take over the legal system.” “What happens to a religion of the book when the greatest expert on the holy book is an AI?”
Meaning requires flesh, not just words: “Human brain development is a product of life experience as sentient being - feeling, loving, anger… And there is maybe something that is still of value there that goes back to that core business of this sentient human being.”
What AI Can Do:
Master all word combinations
Describe emotions perfectly
Generate technically proficient art
Compose music that sounds human
Write poetry that scans correctly
What AI Cannot Do (Per Harari):
Feel the emotions it describes
Experience the qualia of creativity
Create from embodied existence
Generate meaning from suffering/joy/longing
Demis Hassabis acknowledged limits in scientific creativity: “Things like creativity, I think, is still somewhat lacking. So hypothesis generation, coming up with really new things. I think that’s still an area that still needs to be improved.”
Our takeaway:
The silence speaks volumes - both about AI’s creative limits and the market’s likely indifference to quality.
AI’s word capability will improve (brand voice documentation + custom GPT = effective copywriting automation). But as cultural product - creative advertising, music, art, photography - Harari’s words vs. flesh insight is the hard stop.
Expect gigabyte upon gigabyte of trash as human creativity becomes only more precious.
3/ Software is going to be free (or the Enterprise software model severely dented)
“There are going to be basically no human generated lines of code being written by professionals at large companies in 1 to 5 years. That’s a, you know, 5 to 10 million person industry globally or something like that in this one particular area.”
Dario Amodei
Satya Nadella drew a parallel from his Microsoft early days:
“One of Bill’s things at Microsoft from the day I joined in ‘92 always was what’s the real difference between a document, a website, and an application, right? It’s the lack of software that can transform itself.
Interestingly enough, AI finally gives us that, right?
Which is I can write a document. I can just say no, I don’t want it as a document. I want it as a website. It’ll just transform that document using code into a website. I say I don’t like the website. I want an app. It’ll write more code to transform it.”
If AI writes code, why buy enterprise software?
Traditional Model:
Identify need
Buy software
Customise/integrate
Maintain/update
Pay recurring licences
AI Model:
Identify need
Describe to AI
AI generates custom software
AI maintains/updates
Pay for... tokens?
But Jensen Huang maintains this doesn’t lead to the death of the Enterprise software business model. In fact, it means a greater dependence on the most powerful and integrated software at the top of the stack of a five-layer cake:
“At the bottom is energy. The second layer is chips. The next layer above it is the cloud infrastructure. The layer above that is the AI models. But the most important layer... is the application layer above that. This application layer could be in financial services, it could be in healthcare, it could be in manufacturing. This layer on top ultimately is where economic benefit will happen.”
Which implies a new way of creating value with software.
OLD: Sell licences, charge for features, value in functionality
NEW: Provide orchestration, charge for outcomes, value in context + configuration
Example: Salesforce
Before: CRM software with features
After: AI agent that:
Writes follow-up emails
Predicts churn
Recommends next actions
Updates records automatically
Our takeaway:
Two simultaneous trends exist here:
Top layer concentration: Sophisticated software integrates deeper into core verticals (healthcare, finance etc). Companies become more dependent. More sovereignty is handed over. There are fewer opportunities for niche software players.
Bottom layer proliferation: Simple tools created at low cost for long-tail tasks. Simultaneously driving software creativity alongside an ocean of “software slop.”
The trick: If audiences are going to find your tools and stick with them, businesses need better brand, better communications and more media - not massive investments in digital infrastructure.
Coming in Part 2:
Are chatbots commodifying? Or can brands create stickiness?
Will Claude beat OpenAI with B2B clarity vs. B2C chaos?
What happens when niche expertise becomes free?
Why Elon Musk shouldn’t watch your kids. Genuine quote: “Who wouldn’t want a robot to - assuming it’s very safe - watch over your kids.”