Yannick Grenzinger
Product engineer & manager. Engineering Manager @ Payfit. Father of triplets.
Love #psychology, #UX, #product, #DDD and software development in general.
- Appropriate song www.youtube.com/watch?v=4aIY...
- @devoxx.fr quand j'essaie de créer mon compte sur le site du CFP, je me prends une 400 bad request sans plus de détail :( Si vous avez une piste :)
- Bonne année !
- Reposted by Yannick GrenzingerJ'ai décidé aujourd'hui d'utiliser l'intelligence artificielle générative en tant qu'assistant. Un "assistant de pêche à la ligne" pour être précis. Et comme ce sont des machines qui aiment calculer des suites statistiques de mots, on va lui donner ce qu'elle attend et ce pour quoi elle est douée.
- Reposted by Yannick GrenzingerReading is thinking, and afterwards, writing is thinking. Doing an "AI" summary eliminates thinking in both steps. A doc told me some of her colleagues are using "AI" to write their case notes. Her take: oh hell no, because writing the case notes is when you think about the case.
- Le gap entre les délires que Claude Code me fait avec Sonnet 4.5 et les takes éclatés au sol que je lis ou entends (ex: un prompt ... AI engineer peut remplacer toute une équipe) me confirme que la majorité de notre industrie n'a aucune idée de ce que c'est de faire un logiciel facile à maintenir.
- La bonne nouvelle c'est qu'on est pas prêt d'être remplacé. La mauvaise c'est qu'on va devoir bouffer du code de merde pendant encore des décennies.
- Et oui Claude peut générer des milliers de lignes de code rapidement qui sont bonnes à 80-90% ... le problème de la maintenance ont toujours été les 10% qui restent.
-
View full threadOu alors le futur c'est le code jetable 🤔
- Donc MrPhi fait une comparaison erronné entre un ensemble hétérogène d'êtres humains et un LLM et ça va passer crême ... Comme quoi son take down virulent de Luc Julia s'analyse surtout à son émotion et ses biais sur les LLMs. Voir les 8ères minutes ici : www.youtube.com/watch?v=nqFl...
- Reposted by Yannick GrenzingerReductionism can be wrong, but actionable. Complexity can be right, but paralyzing. cutlefish.substack.com/p/tbm-389-ov...
- The gap between AI vendor (or tech influencer or devrel) that sell you that you can now just be an AI manager doing the work for you ... And the unmaintanable code that AI mostly generate is really interesting and tells us a lot why software cost so much at the end ...
- I think I am a simple man I love to build system that just works
- If you booted up a super-smart AI in ancient Greece, fed it all human knowledge, and asked it how to land on the moon, it would respond “You can’t land on the moon. The moon is a god floating in the sky.” How would you get it to realize the moon is actually a big rock?
- That’s a great, poorly defined problem, and I don’t expect AI to solve it anytime soon. www.theseedsofscience.pub/p/why-arent-...
- Why aren't smart people happier? www.theseedsofscience.pub/p/why-arent-...
- I am really bored listening to Conway's law quote. It's like saying again and again that water is wet at some point. The next question is now: why the software boat keeps sinking ?
- China is the future in so many ways
- Reposted by Yannick GrenzingerChina has passed a new law mandating that anyone creating social media content on sensitive topics like medicine, law, education or finance must hold formal qualifications such as a college degree in those fields. Imagine an internet where people actually had to know what they were talking about.
- The Psychology Of Trust In AI: A Guide To Measuring And Designing For User Confidence www.smashingmagazine.com/2025/09/psyc...
- "The idea is that ChatGPT will be your agent, but in reality you are ChatGPT's agent" www.anildash.com//2025/10/22/...
- Reposted by Yannick GrenzingerChatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.
- Reposted by Yannick GrenzingerFrom the same dev: “it’s also that I cannot see management being concerned solving the information visibility that keeps slowing us down. We’re rolling out AI coding tools for devs but no one thinks about the org silos that slow us down 10x more than any coding use case.”
- The LLM and agentic coding revolution allows anyone to start anything. But... It doesn't fix (so far) to maintain on the long term this explosion of codebase and lines of code
- Reposted by Yannick GrenzingerRobot meets bollard. Nothing beats the #WorldBollardAssociation
- "Don’t treat LLM coding agents as highly capable superintelligent systems. Treat them as lazy, intoxicated robots" lnkd.in/eJQhMqVF But fortunately, some have the solution: humans will review the thousands of lines of code generated by LLMs. /s
- "If you learn one thing, learn that shipping is all that matters. If you learn two things, learn that inspiring people is the best way to ship" thechrisperry.substack.com/p/being-a-go...
