Germany de.indymedia
Degitalization: Agility
Germany should become faster in expanding AI; this supposedly requires deregulation and an agile approach. 17 software developers achieved such agility back in 2001. With at the very bottom
Deutsch (original) ·Auto-translated to English
Germany should become faster in expanding AI; this supposedly requires deregulation and an agile approach. 17 software developers achieved such agility back in 2001. With very different results.
Today's degitalization begins with a vibe, with a mood.
The mood of this year's digital summit, which took place in Berlin earlier this week. Daniel and Esther have already described what could be found there: a mixture ofhttps://netzpolitik.org/2026/digitalgipfel-ki-torschnellpanik/">KI-Torsc...andhttps://netzpolitik.org/2026/digitalgipfel-im-takt-der-wirtschaft/">Walzerin economic four-four time, one could attest in short. But that would be a little short-sighted. Because it's actually possible, at least if we tell our Digital Minister Karsten Wildberger and onehttps://bmds.bund.de/aktuelles/reden/detail/eroeffnungsrede-beim-digital...want to believe from Monday to infinitely more.
We are still too slow in the area of so-called artificial intelligence. Courageous decisions are required. Less detailed regulation in advance, more courage to make quick course corrections. So… er, what was it called… agility. Which brings us to a very similar vibe to 2001.
2001, in the mountains of Utah
In 2001, 17 software developers (diversity wasn't particularly high back then) were dissatisfied with their professional work. Processes were cumbersome, more or less all development projects felt like pure bureaucratic tasks, products took years to complete, and in the end they weren't even good.
The Meeting of the Seventeen took place in February 2001 at the Lodge at Snowbird, Utah. Surrounded by mountains, the experts discussed alternatives to project management methods established at the time while skiing and relaxing, with which better digital products could be created in the future. No longer a waterfall, but something different. They agreed on twelve principles based on four core values - thehttps://agilemanifesto.org/">AgileManifesto. The most important core principles require satisfying the customer through early, continuous software delivery, welcoming changes at any time and building projects around motivated individuals in self-organized teams.
The 17 brought with them very different experiences with relatively new project management methods. There washttps://de.wikipedia.org/wiki/Scrum">Scrum, a framework designed to develop projects in short steps, in sprints with continuous customer feedback, to ultimately lead to better end results and happier customers. Then there was also XP –https://de.wikipedia.org/wiki/Extreme_Programming">ExtremeProgramming. A method that focuses on technical excellence through continuous testing and pair programming. There was DSDM (https://en.wikipedia.org/wiki/Dynamic_systems_development_method">DynamicSystems Development Method), which is less about dull processing of requirements and more about requirementshttps://de.wikipedia.org/wiki/MoSCoW-Prioritisierung">MoSCoWprioritized. So requirements that are really important, those that should be, but also those that are dispensable.
With the Agile Manifesto, the 17 developers agreed on a common set of values that would decisively shape the landscape of software development in the years that followed. The dissatisfaction of this group becomes an internationally applicable model for how better digital products can be created in teams.
2026, in an industrial hall in Berlin
In 2026 there will be dissatisfaction in an industrial hall in Berlin similar to that in Utah in 2001. Becausehttps://bmds.bund.de/aktuelles/reden/detail/eroeffnungsrede-beim-digital...of the speechfrom Karsten Wildberger at the digital summit, his wishes sound very similar to those of the 17 developers: not every detail can be regulated in advance. We would have to learn from real evidence and quickly adjust a possible wrong course. If we were to advance artificial intelligence in this way, it could provide us with significantly stronger growth.
–––
Against ID checks online.
We are fighting for an open internet.https://netzpolitik.org/spenden/?via=rss">Thisonly possible with your support.
–––
So more freedom for the makers, the yes-sayers (mainly when it comes to artificial intelligence). Security will also be taken care of, but we don't want to stand in the way of innovation. Innovation is a shy deer that cannot be scared away by too much security at the beginning.
Yes, it is definitely the case that product innovation always requires space to try out new things, to get feedback and also to fail. However, especially in the context of the current discussion of artificial intelligence, it is no longer possible to speak of a trial phase, when often very questionable AI systems are already migrating en masse into more and more areas of everyday life. Be “tried out”.https://netzpolitik.org/2026/behavioursscanner-in-berlin-kreisverkehr-de...Surveillance systems like those at Kottbuser Torin Berlin all day, even if the success is doubtful.
How do better products really come about?
25 years later, perhaps it is time to see whether this call for a better way of working, for more individual freedom for the makers and teams that develop, really leads to better results.
Thanks to scientific support, it is clear that the agility developed in 2001 can have many different consequences for projects and their results. Evil tongues may claim that Scrum only everhttps://www.handelsblatt.com/meinung/kolumnen/arbeitswelt-kann-agiles-ar...projects cannot be completed and they fail all too quickly.https://rim-management.at/agilitaet-unternehmenstreff-metastudie/">Meta...do not come to a clear universal effect of agility, but attest that agile companies are more resilient.
For software, an agile approach using methods such as Scrum does not per se offer an improvement in software quality. But agile methods canhttps://pmc.ncbi.nlm.nih.gov/articles/PMC9486782/">Culturecreate that creates excellent results, for example because good engineering and quality assurance can be practiced.
A culture for good products? This seems to have been completely lost in the fast-moving, wasteful and egocentric times of resource-consuming large language models. Long unresolved problems such as hallucinations, bias or immense need for resources? We may fix it – never. Carefully question the needs of users and consider whether certain AI functions are really helpful? That was yesterday. AI capabilities are now being thrown in everywhere.
In 2022, Cory Doctorow coined the term “https://de.wikipedia.org/wiki/Enshittification">Enshittification". He described the decline of online platforms: First they attract users, then advertising customers, in order to finally exploit both for maximum profit. Agile software development often unintentionally encourages such development. Through incremental, quick updates in sprints, small, profit-oriented deteriorations can be implemented gradually and without much outcry from users, until in the end all that is left is an "enshittifacted" product that was actually quite okay a few releases ago.
–––
The work of netzpolitik.org is financed by donations from our readers.
You too can support our public interest-oriented, advertising- and tracking-free journalism.https://netzpolitik.org/spenden/?via=rss">Donatenow..
–––
Users are rarely happier with the new products, but the economic stakeholders enjoy more profits. Example of a large accumulation of enshittification while optimizing profits at the same time: the transport servicehttps://www.enshittification.xyz/uber">Uber.
New technology does not automatically make everything better, as we are often promised. The emergence of today's omnipresent AI in the form of ChatGPT at the end of November 2022 ultimately brought even more powerful but similarly flawed product releases of chatbots and applications based on them. The emergence of social networks has not led to more digital communities without significant health side effects, algorithms such as TikTok's are consideredhttps://netzpolitik.org/2026/rabbit-hole-effekt-eu-kommission-haelt-tikt...and actually fundamental changes are needed.
Integrated product features
The desire of the economy for more freedom from the “shackles of regulation”, as demanded at the digital summit when dealing with artificial intelligence, is actually more of a reflection of the current demands on the implementation quality of digital products. Who demands data protectionhttps://netzpolitik.org/2026/ki-und-datenschutz-wo-der-digitale-omnibus-...digital omnibusneeds to be regulated differently shows how much he cares about fundamental rights. Anyone who is already demanding that the AI Act must be implemented differently, even though it is hardly in use yet, gives an idea of what the claim is about really wanting to fulfill universal product properties such as AI safety.
According to the agile manifesto of 2001, processes and behaviors were gradually developed in software development, which also included supposedly innovation-hindering disciplines such as security in the product development cycle. Only considering security in software projects towards the end, as Karsten Wildberger indirectly suggests in 2026, is always expensive and hardly leads to truly secure products. This is why agile development has changed in the 2010shttps://eforensicsmag.com/history-of-devsecops/">DevSecOpsestablished an approach that sees security as an integral part of every product development, from the development to the operation of a product. Products that are supported in this safe manner can continue to become safe despite leaps in innovation. But here, security was not seen as an afterthought, but rather as an integral part of a product.
It is usually a result of failed prioritization and often wrong political incentives to want to do something good for the economy through simpler regulation. Certain properties must be required by regulation because the market does not consider them relevant. This was the case with data protection with the General Data Protection Regulation, and it is the same with IT security with thehttps://www.bsi.bund.de/DE/Themen/unternehmen-und-Organisationen/Informa...Resilience ActThis has been the case and this will probably also be the case with products from AI elements in a few years after supposed innovation.
Possibly, just possibly, the last chance for Europe not to miss the boat in the age of so-called artificial intelligence is, sensibly, not to imitate harmful AI products from other countries. Perhaps this opportunity is to create all-round better digital products that really incorporate sustainability, preservation of democracy, compliance with fundamental rights and security into digital product development right from the start. This can also be done in an agile way, but you really have to want it politically and economically.
–––
Your donation for digital freedom rights
We report on current internet policy developments, uncover scandals and initiate debates. We are completely independent. Because our fight for digital freedom rights is financed by donations from our readers.https://netzpolitik.org/spenden/?via=rss">Nowdonate.