iiinova — where your digital self keeps evolving

Valentin Deleplace @deleplace.bsky.social · 27/09/2026

Ce message est tout à fait normal, c'est parce que j'ai un compte #microsoft personnel sans annuaire d'entreprise

$Interaction requise Le portail a rencontré un problème lors de la tentative de récupération des jetons d'accès. Nous vous suggérons de tenter de vous reconnecter ou de continuer sans jetons d'accès, même si cela peut entraîner une expérience utilisateur non optimale. Détails supplémentaires : interaction_required: AADSTS16000: User account '{EUII Hidden}' from identity provider 'live.com' does not exist in tenant 'Microsoft Services' and cannot access the application '74658136-14ec-4630-ad9b-26e160ff0fc6'(ADIbizaUX) in that tenant. The account needs to be added as an external user in the tenant first. Sign out and sign in again with a different Azure Active Directory user account. Trace ID: aac36d8a-f34d-433d-ad64-373002693400 Correlation ID: ffd24d9c-1800-451d-9969-b92ea935f499 Timestamp: 2026-09-27 12:41:22Z. En savoir plus$

000

Valentin Deleplace @deleplace.bsky.social · 07/09/2026

TIL fd -HI you for "I will find you even if you're hiding" 😁 github.com/sharkdp/fd

Usage: fd [OPTIONS] [pattern [path]...]

Options:
-H, --hidden Search hidden files and directories
-I, --no-ignore Do not respect .(git|fd)ignore files

000

Valentin Deleplace @deleplace.bsky.social · 22/08/2026

It's there, ready in the Playground 🥳

000

Valentin Deleplace @deleplace.bsky.social · 15/08/2026

AR24 de Docaposte me propose du papier (inclut dans le prix) ou bien un fichier (payant) 🤦🤦

Preuve de réception
Au format papier (gratuit)
Au format PDF (0.45€ HT)

210

Valentin Deleplace @deleplace.bsky.social · 14/07/2026

La question est bonne et fréquente, mais quelque chose me dit que je ne vais pas trouver la réponse dans la FAQ

000

Valentin Deleplace @deleplace.bsky.social · 08/07/2026

Best $8,000 ever spent on advertising for pushto.computer

000

Valentin Deleplace @deleplace.bsky.social · 02/07/2026

The Push to Computer blog — thirtieth article 🚀 Performance, part 6 — Europe blog.pushto.computer/posts/030-pe... 📱➡️💻

100

Valentin Deleplace @deleplace.bsky.social · 29/05/2026

My time series plot is beautiful like a Pink Floyd album. Smaller is better. The latency values going down at week 21 are excellent news. #BigQuery #SRE #darksideofthemoon

$Time series graph that looks like light being diffracted through a triangle prism$

010

Valentin Deleplace @deleplace.bsky.social · 09/04/2026

Participez au Google Cloud Summit France 2026 cloudonair.withgoogle.com/events/googl... #googlecloud

020

Valentin Deleplace @deleplace.bsky.social · 30/03/2026

Sinon est-ce que Gemini analyse aussi le contenu des images? Oui, absolument:

100

Valentin Deleplace @deleplace.bsky.social · 30/03/2026

Robot, regarde la vidéo à ma place Parmi les meilleurs usages de l'IA, selon moi! Les devoirs de mon fils demandaient "Combien d'adjoints au maire de la ville?" Cette décision a été prise samedi matin au 1er conseil municipal du nouveau maire, et n'était pas encore disponible sur

Un robot qui regarde le conseil municipal à la télé

Dans cette vidéo, combien d'adjoints au maire ont été nommés? https://www.youtube.com/live/MEBNoOjJSGQ?feature=shared

Lors de la séance du conseil municipal du 28 mars 2026, 12 adjoints au maire ont été nommés [57:00]. Voici les détails de cette nomination...

110

Valentin Deleplace @deleplace.bsky.social · 15/03/2026

Loving how Nano Banana 2 multimodal image+text prompt is enhancing my sketch. One-shot 😅😅 #gemini

My sketch+the rendering by Nano Banana 2

111

Valentin Deleplace @deleplace.bsky.social · 11/03/2026

dgraph-io/badger (core) is 5% faster with Go 1.26 #golang dgraph-io.github.io/badger/ go.dev/doc/go1.26

110

Valentin Deleplace @deleplace.bsky.social · 02/03/2026

The Push to Computer blog — twenty-fourth article 🚀 Analytics, part 4 — Location blog.pushto.computer/posts/024-an... 📱➡️💻

000

Valentin Deleplace @deleplace.bsky.social · 27/02/2026

Stellantis lost €25bn in one year, which is more than what it would have cost to make France happy

100

Valentin Deleplace @deleplace.bsky.social · 06/02/2026

L'appli Orange me donne même un schéma assez complet de mes équipements. Well played

000

Valentin Deleplace @deleplace.bsky.social · 06/02/2026

J'ai été agréablement surpris par le répéteur wifi 6 de Orange Très simple d'installation, performant, c'est parfait!

110

Valentin Deleplace @deleplace.bsky.social · 20/01/2026

J'ai cru à un parti politique, mais en fait non

010

Valentin Deleplace @deleplace.bsky.social · 21/11/2025

Antigravity is **super impressive**. I asked it to make a command line client for Cool Maze - Push to Computer, in Go. In 5mn it implemented the complex workflow right, and it **just worked**. Within 2h I had all the less important details right as well. Wow 🤯 antigravity.google #AGY #Gemini

The QR code displayed by the Cool Maze - Push to Computer command line tool generated by Google Antigravity

141

Valentin Deleplace @deleplace.bsky.social · 20/11/2025

A drop in the rating of my mobile app. I guess someone has given 4★. Well, 4.9 now looks less suspicious than 5.0 🤷

The average rating of Cool Maze on Google Play Store dropping from 4,951 to 4,940

020

Valentin Deleplace @deleplace.bsky.social · 07/10/2025

Linkedin knows me so well 🤣😂 Thank you AI and big data

Valentin, explore relevant opportunities with PMU

Get the latest jobs and industry news

010

Valentin Deleplace @deleplace.bsky.social · 08/09/2025

Hmm what's this spike at 18 Cloud Run instances? A traffic surge? An infinite loop in my backend code? An attack? A misconfigured autoscaler? #GoogleCloud

Container instance count chart: always a single instance, except at 8:40PM with 18 instances

100

Valentin Deleplace @deleplace.bsky.social · 17/08/2025

Microsoft trying EEE "Embrace, Extend, Extinguish" to fight spam or something

Suspicious but genuine email saying "Bonjour Valentin! Achetez avec Microsoft et gagnez de l'argent en retour"

000

Valentin Deleplace @deleplace.bsky.social · 31/07/2025

In other shocking news, the "fastest" way to send a photo from mobile to a nearby computer is effectively to upload it to a server located in another continent, and have the computer download it from that distant server 👇

John Carmack: "I can send an IP packet to Europe faster than I can send a pixel to the screen. How f’d up is that?"

100

Valentin Deleplace @deleplace.bsky.social · 20/07/2025

It it surprisingly cumbersome to determine the Error Correction Level (ECL) of a given QR code, as QR scan libraries don't expose this low-level info. TIL I can just visually read it 🤦 which was enough for my specific needs

How to determine the ECL L, M, Q or H just by looking at the bottom-left of the QR code

010

Valentin Deleplace @deleplace.bsky.social · 23/06/2025

I was horrified reading this. In less than 10s, I had formed a plausible explanation like "selection bias", "Simpson's paradox" or "Berkson's fallacy". How could this puzzle remain unsolved for 40 years? Was the 20th century scientific culture so different? That's the real puzzle IMO

SMOKING FOR NEWBORNS

Even after the smoking and cancer debate died down, one major paradox lingered. In the mid-1960s, Jacob Yerushalmy pointed out that a mother's smoking during pregnancy seemed to benefit the health of her newborn baby, if the baby happened to be born underweight. This puzzle, called the birth-weight paradox, flew in the face of the emerging medical consensus about smoking and was not satisfactorily explained until 2006—more than forty years after Yerushalmy's original paper. I am absolutely convinced it took so long because the language of causality was not available from 1960 to 1990.

In 1959, Yerushalmy had launched a long-term public health study that collected pre- and postnatal data on more than 15,000 children in the San Francisco Bay Area. The data included information on mothers' smoking habits, as well as the birth weights and mortality rates of their babies in the first month of life.

110

Valentin Deleplace @deleplace.bsky.social · 20/06/2025

J'ai été horrifié par ce passage. En moins de 10s, j'avais formé une explication plausible de type "biais de sélection" ou "paradoxe de simpson". Comment cette énigme a-t-elle pu perdurer 40 ans? La culture scientifique du 20e siècle était-elle si différente? Là est plutôt la vraie énigme, selon moi

LE TABAGISME CHEZ LES NOUVEAU-NÉS

Même après la fin du débat sur le lien entre tabagisme et cancer, un paradoxe majeur perdura. Au milieu des années 1960, Jacob Yerushalmy fit remarquer un fait singulier : la santé des nourrissons qui naissaient en sous-poids était meilleure si leur mère avait fumé pendant la grossesse. Cette énigme, appelée « paradoxe du poids de naissance », allait à l'encontre du consensus médical émergeant sur le tabagisme et ne fut expliquée de manière satisfaisante qu'en 2006, soit plus de quarante ans après l'article de Yerushalmy. S'il fallut un tel délai pour résoudre ce paradoxe, c'est parce que – j'en suis absolument convaincu – le vocabulaire de la causalité n'était pas encore disponible entre 1960 et 1990.

En 1959, Yerushalmy lança une étude de santé publique censée recueillir à long terme des données pré- et postnatales sur plus de 15 000 enfants de la Baie de San Francisco. Les données incluaient des informations sur les habitudes tabagiques des mères, ainsi que sur le poids à la naissance et le taux de mortalité des bébés au cours du premier mois de vie.

100

Valentin Deleplace @deleplace.bsky.social · 28/05/2025

I like that when a mobile app's rating is high, the vertical axis has a 5.5 tick mark

The chart of Google Play ratings for my app, with a vertical axis that goes from 4.5 to 5.5

000

Valentin Deleplace @deleplace.bsky.social · 23/05/2025

😱 HOW did I not know yet hitting "?" in #BigQuery would let me see and edit the keyboard shortcuts. So... New query tab is: ⌘ Alt T

000

Valentin Deleplace @deleplace.bsky.social · 07/04/2025

False positive alert, but I'm thankful these guardrails exist

mail.google.com says

It seems like you have forgotten to attach a file.

You wrote "is attached" in your message, but there are no files attached. Send anyway?

000

Valentin Deleplace @deleplace.bsky.social · 04/04/2025

Dark mode in #GoogleCloud ? 😱😱 I've been waiting for this for 13 years, and it's there now!! The setting is at console.cloud.google.com/user-prefere...

Dark mode Public Preview is now available in certain products

BigQuery UI in dark mode, black background, on a mobile device

100

Valentin Deleplace @deleplace.bsky.social · 21/03/2025

This is important: is the new Go 1.24 benchmark loop syntax making my code 7x slower, and why? #golang

func mult(a, b float64) float64 {
return a * b
}

func Benchmark1(b *testing.B) {
for i := 0; i < b.N; i++ {
mult(3.0, 5.0)
}
}

func Benchmark2(b *testing.B) {
for range b.N {
mult(3.0, 5.0)
}
}

func Benchmark3(b *testing.B) {
for b.Loop() {
mult(3.0, 5.0)
}
}

1206

Valentin Deleplace @deleplace.bsky.social · 10/03/2025

OpenCore Legacy Patcher literally saved me $1500 so it's fine I give them $40 "Install recent macOS on unsupported Macs" github.com/dortania/Ope... opencollective.com/opencore-leg...

The OpenCore Legacy Patcher contribution form

Thank you! You are now supporting OpenCore Legacy Patcher.

000

Valentin Deleplace @deleplace.bsky.social · 16/02/2025

When sharing with Cool Maze, by default photos are resized to max 2000x1500. This is why you have an explicit button "Send original", to transfer the full-resolution original resource

100

Valentin Deleplace @deleplace.bsky.social · 14/02/2025

Oh wow, Cool Maze rated 4.8⭐️ on both the Play Store and the Apple Store coolmaze.io

041

Valentin Deleplace @deleplace.bsky.social · 10/01/2025

Squiggly yellow line under "yield"? Yes, this is a very welcome gopls check for incorrect iterator implementation github.com/golang/go/is... #golang

$// Returns an iterator that produces only the odd // values of a, and discards the even values. func odd(a []int) iter.Seq[int] { return func(yield func(int) bool) { for _, v := range a { if isOdd(v) { yield(v) } } } }$ yield may be called again after returning false

yield may be called again after returning false

100