The Last Shreds of Value as an Information Worker
In this Post I ponder the last shreds of value that Information Workers can provide as AI violently replaces the value of humans.
Compression of Value
Value changes over time and with time comes different cycles, each completing faster than the one before it. Value went from the Agrarian Age (thousands of years up to ~1760) to making physical products during the Industrial Age (~1760 - Onwards) to value being manifested through the capture, flow and analysis of information in the Information Age (~1950 - Onwards).
1000s years → 200 years → 70 years
What is considered valuable shifts in eras, getting ever shorter.
We are now seeing a radical decline in the Information Age. The utility of Information Workers is no more. Replaced by AI.
AI Wins on Knowledge
White collared workers were paid for their value for a few reasons.
First of all there was knowledge. Years of school, university and professional training got loaded into individuals to be replayed on demand. The ability to recall knowledge was valuable.
Transformer Models afforded AI the ability to store knowledge from ingesting books, video and audio, meaning that there was now an artificial way to store and recall knowledge.
Suddenly the need for recall from a human was not-needed. And too with that comes the negation of need for higher education and professional education.
For my role as a Software Engineer - I used to need to know a universe of technology, now I don’t need to know anything except for the core principles (a few rules of thumb that lead to good software). Software developers would have to learn multiple new technologies every month. Now - there is no point learning any technology. You don’t write technology any more. You use the AI to write it for you. You should be re-focussing any time learning into a new field where value still exists.
The idea of an individual taking on hundreds of thousands of debt to gain knowledge is truly idiotic in 2026. How many years in a Post-Tax earnings environment do you need to earn to get back from 100k to zero?
AI Wins on Logic
The knowledge that the GPT’s (General Purpose Transformer) crystalized isn’t rigidly stored as isolated nuggets. There is a layer of logic on top of if that allows other topics to be imbued on top.
Any number of topics can be blended together in meaningful and pragmatic information extraction.
If a human’s value production was similar to a gemstone, it would have many facets on its surface. Slowly the facets are disappearing from how information workers produce value.
AI has reasoning logic that can outpace a human in most scenarios. Its ability to extract knowledge and apply logic on top has made it highly valuable.
AI Wins on Execution
With the addition of the creation of agentic behavior, generic AI can call other task specific AIs to cause an outcome. This is Agentic AI.
Generally speaking - AI forms into a tree structure, the roots or trunk are more generic and higher cost thinking. They are the Architects/Designers and delegate down to branch and leaf nodes. The closer you get to a leaf node, the more purpose built the AI tooling is and also the dumber/cheaper.
Somewhat modelling an organization - High level decisions and accountability at the top then delegating down to hierercies of specialization.
AI can be now left by itself to design, review, create.
The value from executing has disappeared as a facet of value for the information worker too.
Execution is both in the forms of throughput and Creativitiy. Images generated by AI are better than most humans can do.
What Facets Remain of Value for the Information Worker
Clearly information workers still provide some value as they are still being employed, but it is expected that their throughput should have multiplied. I know for a fact im outputting somewhere up to an order of magnitude more output than I was 3 years ago.
This increase in throughput comes with the need for the individual to shift from single task implementor to one more of a Project Manager (of agents) or multitasker.
This comes with a real requirement of physiological change. Decades of neuroplasticicity from similar work suddenly needs to shift into a different way of thinking - the real need for concurrency in thinking.
Personally I have adapted by using each agent to track its own progress against a pre-established implementation plan. That way it is quick and easy to check the progress of the multiple contexts going on at once. This alleviates your cognitive burden.
However I don’t suppose that this facet of “throughput” will exist long given that AI can improve exponentially.
Devloping Emotional Intelligence, body language awareness and the ability to speak to external clients in a support role or a sales role are definately an area of value. This is an area where humans can’t be replaced so clearly an area one must double down on.
I’ve actively started to shift into this. Taking my knowledge of Software, but using it to talk straight to the client or to act as a Sales representative.
Humans Still Win on Agency
Agency is a trait that only actual living lifeforms possess right now.
Getting out of bed and walking down stairs. Randomly insulting someone for the hell of it. Setting goals for the new year.
Agency would be deciding then action upon something.
Not from programming. Not from instruction by a human - genuine self driven agency.
It is an independent decision to change and affect the surrounding universe.
For humans providing value. this might take the shape as being a thought leader, or an entrepreneur who seems a gap, decides to act upon it and produce a product. For science or math, agency would be going out and finding new Math, new Science that didn’t exist before, without being prompted to do so.
To my knowledge as of 2026 - AI does not have agency to go and create new knowledge/science/products that are beneficial, without being prompted.
If you form the snowball with your own hands and roll it down the hill, AI will take over and make that snowball very big. But the first few steps do not exist yet.
Humans Still Win on Accountability
Accountability cannot be delegated to AI as humans don’t value AI to the same level as a human yet.
Accountability for large decisions lie with senior management of companies. It would be quite hard to imagine a time where Accountability can be put square on the shoulders of AI.
Humans Still Win on Physical Interaction
When it comes down to humans interacting with humans - humans win with regards to empathy as well as fiddly ad-hoc manipulation of physical things. So for the time being - blue collared jobs are safe.
Human to human interaction will always be preferable for healthcare and sales.
Conclusion
The number of facets of value that humans can provide as an Information Worker are disappaering extremely quickly. Governments are accelerating this anihilation of the status-quo through printing money and firehosing it straight into AI Data Centers.
When making forward looking decisions, one should not base them on familiarity with the past and present. Generally speaking for a global population, the remenant value is situated in the Physical world of things plus the Attention Economy (Being a creator).
The White Collared Work ship is sinking and all those aboard should make plans to pivot.
Can you think of any facets of value that I haven’t covered - please add them as comments if you can think of any!