With AIs wide open…

A presentation at All Things Open in October 2025 in Raleigh, NC, USA by Chris Heilmann

Slide 1

Slide 1

With AIs wide open… All Things Open 2025 Chris Heilmann, WeAreDevelopers

Slide 2

Slide 2

30 years of sneaking ”open” into my work… • • • • • • • • • • • Radio newscaster => quit, taught myself HTML and Perl Agency work eCommerce Etoys.com - UK, US DotCom crash Government websites, larger clients Enterprise CMS (India, UK) Yahoo (yay open!) Mozilla (even more open!) Microsoft (lots of open) WeAreDevelopers

Slide 3

Slide 3

WeAreDevelopers = Europe’s biggest independent developer community Newsletter Podcast Live Show Magazine

Slide 4

Slide 4

Slide 5

Slide 5

“Open” means… • Access for all, independent of location or wealth • Access to functionality (APIs) with documentation • Access to code • In-built quality control (lots of eyes) • Longevity – people can build on your work • Building things that work • Following agreed procedures to ensure quality The current hype around AI is the antithesis of almost all of that.

Slide 6

Slide 6

AI overviews instead of search results = the truth according to company X

Slide 7

Slide 7

AI Image and video generation = expensive, faux creativity without license considerations

Slide 8

Slide 8

Chat systems and research helpers = unsolicited scraping and bots serving bots

Slide 9

Slide 9

Agents browsing and buying things… = a malware creator’s wet dream

Slide 10

Slide 10

AI aided development • “AI” IDE = fork of Visual Studio Code with all security settings turned off. • AI features of these “free tools” costs you tokens • On average, this means $100k for a development team/year. = subscription in disguise

Slide 11

Slide 11

Slide 12

Slide 12

Slide 13

Slide 13

Slide 14

Slide 14

Don’t trust the bots… • Only 33% of developers trust AI accuracy in 2025, down from 43% in 2024 and 42% in 2023. • AI favourability dropped from 77% in 2023 to 72% in 2024 to just 60% this year.

Slide 15

Slide 15

Fixing bad results isn’t fun… 66% of developers cite “AI solutions that are almost right, but not quite” as their top frustration 45% say debugging AI-generated code takes more time than expected. AI tools promise productivity gains but may actually create new categories of technical debt.

Slide 16

Slide 16

Writing or generating code should always aim to solve an issue. How much code we write in what time is a pointless metric…

Slide 17

Slide 17

Vibe coding

Slide 18

Slide 18

Slide 19

Slide 19

C

Slide 20

Slide 20

CEOs can now do production code! Our first outage from LLM-written code https://sketch.dev/blog/our-first-outage-from-llm-writtencode?utm_source=tldrnewsletter

Slide 21

Slide 21

Finding the issue… “The trigger for the CPU spikes both times was our CEO logging in. We re-deployed to get a clean start, permanently banned him from the service, and moved on.”

Slide 22

Slide 22

Slide 23

Slide 23

Slide 24

Slide 24

Vibe coding full products… = dubious, insecure legacy code

Slide 25

Slide 25

Let’s talk about we use AI…

Slide 26

Slide 26

We have a ton of content! • Weekly newsletter with 150k subscribers • Weekly live show with guests • Magazine articles every few days • Podcast every few days • 10 years WeAreDevelopers World Congress: • ~2500 – 3000 hours of talk and workshop footage x 10 • This year it is 12 TB of data

Slide 27

Slide 27

Slide 28

Slide 28

Indexing and remixing > Generating with AI

Slide 29

Slide 29

Work in progress…

Slide 30

Slide 30

Using AI to make video more accessible…

Slide 31

Slide 31

Your dev skill playlist…

Slide 32

Slide 32

Let’s talk about our jobs, longevity and how we deliver software.

Slide 33

Slide 33

Let’s talk about our jobs, longevity and how we deliver software.

Slide 34

Slide 34

Let’s talk about our jobs, longevity and how we deliver software.

Slide 35

Slide 35

Let’s talk about our jobs, longevity and how we deliver software.

Slide 36

Slide 36

We learn by imitation and from mistakes. How can junior developers learn if they’re encouraged to let magical AI do everything for them?

Slide 37

Slide 37

Dunning-Kruger Effect A cognitive bias, where people with little expertise or ability assume they have superior expertise or ability. This overestimation occurs as a result of the fact that they don’t have enough knowledge to know they don’t have enough knowledge. … we made this bias the go-to effect of AI …

Slide 38

Slide 38

Slide 39

Slide 39

We protected the web from corporate takeover before. We need to do this again.

Slide 40

Slide 40

Not by demonising AI and the companies that succeed in it. But by taking part and offering alternatives…

Slide 41

Slide 41

AI can make us more efficient; it can’t make us more creative or magically excellent at anything.

Slide 42

Slide 42

Let’s show the world how the power, love and dedication of open can move this from marketing hype to the next level of human-computer interaction.

Slide 43

Slide 43

Free WeAreDevelopers World Congress Berlin, Germany 08-10/07/2026 (if you get picked as a speaker)

Slide 44

Slide 44

WeAreDevelopers World Congress North America San Jose, California 23-25/09/2026 10% cheaper! Free (if you get picked as a speaker)

Slide 45

Slide 45

Thanks! ChristianHeilmann.com WeAreDevelopers.com Developer-Advocacy.com chris@wearedevelopers.com @codepo8@toot.cafe @codepo8.bsky.social linkedin.com/in/christianheilmann/