In a session today there was some discussion about agentic skills. A skill is mostly a markdown text file often named, yes, skill.md with a natural language text plus some front matter, describing some specific workflow or bit of knowledge or capability.
Somebody external opined that such skills were a good way of codifying organisational knowledge, which might be useful ‘because of the ageing demographics’ on the work floor and people retiring.

Would agentic skills succeed where process descriptions, intranets, 1990-2010 knowledge systems, quality assurance systems all failed when aimed at documenting the actual knowledge present in an organisation?

I doubt it.

Deep knowledge is relational, chunky, only its underlying facts and heuristics can be codified. And people will resist codification, for different reasons. Such as recognising the futility of it, or recognising its aim of making the people with knowledge superfluous. Or both.

Are skills different enough from earlier systems aimed at codification to gain some willing cooperation from knowledge workers?

They might be.

Unlike the earlier systems, a skill is small and its aim is not explicitly ‘codifying everything’. Smaller than you, an information object in itself. And you get to combine such objects into a chain or bundle as a small workflow. They are a much smaller ask than the others were.
Additionally skills can be deployed by yourself in your own work. For processes, QA, KM it was mostly all done for others, for the hand-offs between actors and mostly seen as external to one’s own work. Why would you write down and maintain what you would not use yourself?
That is where skills can have a different impact. And in terms of ease of use, it helps that such skills are in natural language, as it allows you to handwave the parts where you can’t express your own actual knowledge in any precise detail. Which will be often.

So skills will be likely an easier sell to do some codification.
Still the push for skills is to be small, niche and specific. I’ve seen lots that are actually deterministic, in other words can be automated, except perhaps for the decision when to call a skill and with which inputs. Meaning they codify something different than knowledge.
Such codification will not likely form a coherent whole either. Also because skills are regularly aimed at something that is not within the scope of existing knowledge, a helper function next to current knowledge work.

So, no, I don’t buy the ‘skill codification as knowledge transfer’ suggestion to sell skills with.

In a session earlier today the deployment of a tool was discussed in an organisation’s information environment that is otherwise fully dominated by Microsoft products. After a presentation a director asked “if we add this tool to our (Microsoft) mix, aren’t we working in opposition to our wish for more digital autonomy and digital sovereignty”? The answer was “well you are already using Microsoft tools everywhere so sourcing this one tool from someplace else won’t change your autonomy in any way.”

I disagree. This is the pitfall of “it’s only worth changing if we can get from here to a mythical ideal situation in a single easy step, otherwise we’ll stay put”.
Indeed, opting to source a single tool outside of the walled garden you’re in does not change anything about you being in that walled garden as such. But adding a tool within the walled garden does make it harder still to leave that walled garden in the future. It increases your cost of leaving once more.

You have to begin somewhere if you want to move out of monopolies under foreign jurisdiction outside your democratic control, that can determine what’s on your screen and in your inboxes. Not adding to the problem is one step easily made, because it does not require you to change anything in your current way of working, just to apply different decision criteria upfront when adding something.

Bookmarked How Generative and Agentic AI Shift Concern from Technical Debt to Cognitive Debt by Margaret-Anne Storey

I enjoyed this short posting by Margaret-Anne Storey, a CS professor. The effect of using generative tools can indeed lead to loss of overview, and uncertainty about the project I recognise. It creeps in very quickly, especially if I’ve started from something exploratory, as opposed to planned. A cognitive debt accrues because of wanting to move fast or move at all, at the cost of understanding one’s actions in enough detail. It hinders being able to make changes later.

It also makes me wonder something completely different. Partially because of examples I saw last week in Madrid of how BMW and Airbus had sped up some specific tasks orders of magnitude with AI:

If we see companies as slow AI, i.e. context blind algorithms working towards a narrowly defined singular goal (this is where the notion of AI turning all the material in the world including ourselves into paperclips comes from), what methods have we come up with to deal with cognitive debt in organisations? My intuitive response is reporting chains, KPIs, and middle management. Consultancy too, hiring an external actor to blame if needed. That suggests to me we actually didn’t, as so much of that is management-theater. Does any board of any company above a certain size actually know what is going on in their organisations? Understand what consequences changes may have? There’s a world of hurt out there caused by ‘reorganisations’ that all too often seem ritualistic more than rational when seen from the outside.

It may also be why companies easily embrace AI, despite e.g. warnings about cognitive debt. It looks the same as current practice, just with the promise of higher speed.

I saw this dynamic play out vividly in an entrepreneurship course I taught recently. .. one team hit a wall. They could no longer make even simple changes without breaking something unexpected. … no one on the team could explain why certain design decisions had been made or how different parts of the system were supposed to work together. … issue was that the theory of the system, their shared understanding, had fragmented or disappeared entirely. They had accumulated cognitive debt faster than technical debt, and it paralyzed them.

Margaret-Anne Storey

Twenty years ago today E and I visited Reboot 7 in Copenhagen. What I wrote a decade ago at the 10th anniversary of that conference still holds true for me.

Over time Reboot 7 became mythical. A myth that can’t return. But one we were part of, participated in and shaped.
Still got the t-shirt.


The yellow t-shirt with red text from the 2005 Reboot 7 conference, on my blue reading chair in my home office 20 years on.

Seventeen years ago today I blogged about a barcamp style event in Amsterdam I co-hosted, called GovCamp_NL. I struck up a conversation there about open government data after having had a similar conversation the week before in Austria. It marked the beginning of my work in this field. We just welcomed the thirteenth team member in the company that over time grew out of that first conversation. Our work at my company is driven by the same thing as the event, something I’ve come to call constructive activism.

These days, the principles and values that drove those events, and have set the tone for the past two decades of everything I’ve done professionally and socially, seem more important than ever. They are elemental in the current geopolitical landscape around everything digital and data. We can look back on our past selves with 20 years hindsight and smile about our one time optimism, because so much exploitation, abuse and surveillance grew out of the platforms and applications that originate in the early 00’s. But not because that optimism was wrong. Naive yes, in thinking that the tech would all take care of itself, by design and by default, and we just needed to nudge it a bit. That optimism in the potential for (networked) agency, for transparency, for inclusion, for diversity, and for global connectedness is still very much warranted, as a celebration of human creativity, of the sense of wonder that wielding complexity for mutual benefit provides, just not singularly attached to the tech involved.
Anything digital is political. The optimism is highly political too.

The time to shape the open web and digital ethics is now, is every day. Time for a reboot.

I notice a strong and persistent reluctance with Dutch civil servants to use the word citizen. Apparantly because the Dutch word ‘burger’ carries overtones of ‘kleinburgerlijk’, petty bourgeois, of bourgeoisie, and of the general disdain university students voice for ‘burgers’ (with ‘burger’ being bandied about as an insult amongst them, which gained national usage through the 1990’s Jiskefet satirical tv program). Many civil servants said to me they think the word citizen is ‘old fashioned’.

I find this not only an oddity, but also detrimental to public governance and potentially dangerous.
Not using the word citizen obscures how in the relationship to government citizens have basic human rights, specific constitutional rights, and some duties. A citizen has autonomy and a certain power vis-a-vis the government.
Not using the word citizen, easily obscures that power and those rights to civil servants.

I hear civil servants talk about

  • ‘customers’, usually in the context of providing public service
  • ‘clients’, often in the context of the social domain, reminiscent of how therapists talk
  • ‘inhabitants’, usually a hand-wavy acknowledgement that other people are involved, but in an abstracted, passive or even statistical way,
  • ‘users’, usually carried over from an IT related context
  • or worst case ‘residents’ as if you’re institutionalised.

In all these cases it creates either a distance to people or implies power assymmetries. It makes it easier to dehumanise people. The consequence is the creation of policies about people, but not with those people, because people are never perceived to be on equal footing. Policy gets done over people’s heads, done to them. Participatory processes are then easily reduced to a ritual, a checkbox to mark, something that is a pain and a drag without which your policy process would be so much more efficient. Clients, users and inhabitants are never equal to those who determine policies, whereas citizens would have to be met eye to eye. Acknowledging people as citizens would require curiosity about their needs, motives and actual experiences when developing policy.

Every civil servant I’ve worked with cares about good governance and public service, and individually they wouldn’t treat people as passive objects on which their policies operate, but collectively in their work context they do abstract people out of the equation. And their own choice of words contributes to that, makes it more likely to happen, I think.

In conversations with our public sector clients I always talk about citizens with emphasis. I often also introduce myself as citizen (not as consultant e.g.).

In our projects we always emphasize the need for civil servants to go outside, to check their data and documents against the reality outside, and as often as possible create conversations with real people, with citizens.

With the drive towards ‘data driven’ work, this is ever more essential. Data must be presumed to always describe only a sliver of reality, and to always do so badly on top of that. There is always a check against reality necessary when you want to start relying on data in policy decisions. Visit the places and the people represented in the data, do you recognise them? Do you have a sufficiently nuanced, detailed and rich view on an issue before making a decision? Do people’s stories validate the data, is their meaning incorporated?
Acknowledging people as citizens is also essential to being able to see and use government data publication as a policy instrument, meant to provide agency to people in the context of societal issues and as equal partners in addressing these issues.

Hight time for the public sector to use the word citizen routinely and meaningfully again.

Favorited How Big Tech Runs Tech Projects and the Curious Absence of Scrum by Gergely Orosz

I see scrum used at different levels of quality in different settings, and especially where the focus shifts from delivery to the rituals around enabling the delivery it becomes a drag quickly. Then there’s the ever increasing backlogs and the partial implementation where scrum delivery collides with the rest of an organisation not being ready to pick up on what’s delivered. (HT to Alper Çuğun for surfacing the link in his rss feed)

The success of companies and project management approaches is not always correlated and this story is a reminder of this…..the organizational structure of Big Tech greatly impacts how teams can – and do – execute. …. When talking to engineers at Facebook, Whatsapp, Google, Netflix and similar organizations, most of them have never used Scrum. Should companies dismiss Scrum and other methodologies just because most of Big Tech has done so? … There are many contexts in which switching to Scrum makes perfect sense

Gergely Orosz