Marc Englund (@marcenglund) header
Marc Englund (@marcenglund) profile picture

Code and photos, professionally and for fun. Product Strategist @Vaadin

Jakobstad, Finland
Joined April 2008
Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)Photo by Marc Englund (@marcenglund)
You would like to do what? Seems a bit excessive, no? Especially since the update completed fine w/o granting this. Seriously WTF @ASUS
44
Marc Englund retweeted
After a few months of intensive work with AI agents on Java projects, I can say: Vaadin is currently the best web framework for Java developers. Two reasons make the difference. ๐Ÿญ) ๐—•๐—ฟ๐—ผ๐˜„๐˜€๐—ฒ๐—ฟ๐—น๐—ฒ๐˜€๐˜€ ๐—ง๐—ฒ๐˜€๐˜๐—ถ๐—ป๐—ด Karibu Testing and the official UIUnitTest let me test complete views without a browser. No WebDriver, no flakiness, no 30 seconds wait per run. Tests run in milliseconds. This sounds like a pure comfort argument, but it is much more. When I let an AI agent work on a Vaadin project, the test suite runs after every change. The agent sees right away if it broke something and fixes it. With a slow E2E suite, this does not happen. The agent gets the feedback too late, and quality drops. The Test Trophy model fits very well here: the focus is on integration tests through the whole view, not on isolated unit tests of single components. ๐Ÿฎ) ๐—”๐—œ ๐—ฆ๐—ธ๐—ถ๐—น๐—น๐˜€ ๐—ฎ๐—ป๐—ฑ ๐— ๐—–๐—ฃ ๐—ฎ๐—ฟ๐—ฒ ๐˜๐—ต๐—ฒ๐—ฟ๐—ฒ Vaadin has released an official Claude Code plugin: github.com/vaadin/claude-pluโ€ฆ It comes with eleven curated skills (Theming, Layouts, Forms, Signals, Views, Testing and more) plus an integrated Vaadin MCP server for live documentation lookups. The effect in daily work: without MCP, the agent guesses from its training data. That data is months old and does not know the current Vaadin 25 APIs. The result: hallucinated methods, outdated patterns, code that does not compile. With MCP, Claude queries the official documentation live. With the skills, the agent gets curated best practice knowledge on top, which is not part of the general training data. The difference in code quality is clear. ๐—ง๐—ต๐—ฒ ๐—ฏ๐—ถ๐—ด๐—ด๐—ฒ๐—ฟ ๐—ฝ๐—ถ๐—ฐ๐˜๐˜‚๐—ฟ๐—ฒ I believe that frameworks will no longer be measured only by their Developer Experience, but also by their Agent Experience. Short feedback loops, curated knowledge via MCP and skills, consistent patterns: these are the new quality criteria. Vaadin already meets these points today. Java instead of JavaScript, server-side, fast tests, official AI tools. Exactly the combination that matters in the enterprise Java space. This aligns strongly with what I teach in my SDD Workshop and in the AIUP methodology: anyone working with AI agents needs the right tooling around them, otherwise you fight against the agent instead of with it. #Vaadin #Java #SpringBoot #AI #ClaudeCode
๐Ÿค– Made with AI
1
6
30
1,362
Restarting your http MCP server? โœ”๏ธCodex CLI reconnects (JIT connect, I think) โœ”๏ธ Claude Code started reconnecting "recently" โŒ Claude Desktop via mcp-remote does not reconnect Other findings/corrections? Is there a cross-agent feature matrix somewhere?
1
170
Did you know @vaadin 24.8 now offers a super-simple way to share feedback? ๐Ÿš€ Give the Beta a spin and tell us what you think (how meta)!
3
3
491
Watch this if you're even slightly interested in how AI powered apps are built! @marcushellberg does a great job showing the most important bits and illustrating what they actually do in an approachable manner โ€“ even if you're not a @java @springboot @vaadin type of person.
By popular demand: I published a live code demo and tutorial on how to build your own AI powered assistant with @springboot and @langchain4j with a @reactjs frontend powered by the @vaadin Hilla framework: youtu.be/NL7bqlTHjtE I've given this tutorial as a presentation at tens of events around the world and I've heard back from many attendees that they've been able to go and build their own cool apps after attending the session. Hope you find it useful!
1
3
379
Hey @vs_marketplace, was support replaced by an AI, or what's going on? - We were already verified back in April. - I resubmitted and replied last week; no response. - What does "by date based on publisher date 10/04/2024" exactly? - Does "by" mean "after" in your language?
1
82
Join @leifastrand @marcenglund and me for a @vaadin 24.4 demo on Tue (8am PT / 6pm EET). Bring your questions! youtube.com/watch?v=J1HH4UIeโ€ฆ
1
3
189
Marc Englund retweeted
๐Ÿ“ฃ SAVE THE DATE! Vaadin Create conference returns on 29-30 October 2024 in Frankfurt, Germany! ๐ŸŽ‰ Join us for two days of innovation, learning, and networking with the #Vaadin community. Stay tuned for agenda and registration details. We hope to see you!
10
3
14
2,562
โ€œDave is deaf, and Wally is blind. They witness a murder, but it was Dave who was looking at her, and Wally who was listening.โ€ Multimodal often reminds me of โ€œSee No Evil, Hear No Evilโ€ (1989, viewer discretion advised), but #OpenAI & #GoogleAI look really promising now!
1
64
Friday glitch or a new type of #MFAFatigue attack? ๐Ÿค” Safari asking for Touch ID, but Iโ€™m not in Safari, and I canโ€™t find which page is asking. โ€œCancelโ€ does nothing. Seriously: Touch ID should never be active w/o the target being visible and focused on screen.
2
207
A sneak peek of my next Java AI business app starter for analyzing review sentiments and drafting custom responses (using LangChain4j and Hilla). I want to create a handful of these starters: what are some other use cases for AI in business apps that you'd like to see?
2
4
23
3,417
I spoke about โ€œAI-powered business applicationsโ€, why itโ€™s crazy right now but still the perfect time to act + some examples. As itโ€™s about AI you should watch it right now, because tomorrow it might be outdated: youtu.be/J3wbPLrxpvE?featureโ€ฆ
49
Hey @builderio, I just noticed all the #opensource links seem to go to the wrong place. Been there, done that, thought I'd let you know ๐Ÿ˜‰
1
86
The subscription experience for @OpenAI must be one of the most confusing and buggy ones I know. Others tend to make it easy to keep paying. Maybe ChatGPT is coding its own site โ€“ that would explain a lot ๐Ÿ˜‰
62
I mentioned to @LawrenceDCodes that I wanted to show him how I like to build @hillaframework apps and share some tips and tricks. We decided to live stream it on YT so everyone can join. What could go wrong? Join us on Tuesday and find out! ๐Ÿ˜„ youtube.com/watch?v=yPv7U3Pgโ€ฆ
2
1
13
2,748
Today, #JugOberpfalz hosts @MattiTahvonen ๐Ÿ‡ซ๐Ÿ‡ฎ! Matti is a long-time open-source veteran at @vaadin. In his talk about "You are not Google" he will explore if "best practices" from Big-Players like Google work for "normal companies." meetup.com/de-DE/jug-oberpfaโ€ฆ
3
13
1,206
I got a ton of questions on how I built the @hillaframework Docs Assistant ๐Ÿค– so I wrote a blog post outlining the process: marcushellberg.dev/how-to-buโ€ฆ tl;dr: create vector embeddings for docs sections, include the most relevant sections as context for ChatGPT
2
6
461
If at all possible, try to avoid this situation. ...also, the maximum number of virtual screens on MacOS is 16. If there needs to be a limit, I feel it should be 13 and increase with each major. Either that, or 42.
1
1
208