Google May Have To Share Search Data With Rivals
The European Commission proposed Google share data with rival search engines and qualifying AI chatbots in the EU/EEA.
#google #seo
Posts by michel verstrepen
🎧 Episode 366 of the Easy EdTech Podcast is LIVE #EduSky
🎯 I share five NotebookLM features that might change how you teach.
🔗 Tune in for practical (and a little wild) ideas you can use right away. classtechtips.com/2026/04/14/n...
Google Gemini gets a productivity boost with synced 'Notebooks'! Organize chats, upload sources (PDFs, docs), & set custom instructions. Syncs with NotebookLM. Available now for Google AI Ultra/Pro/Plus subscribers on the web. #GoogleAI #News
This week in The AI Vaults: Google merges NotebookLM into Gemini, three rival labs team up against Chinese model theft, and Anthropic shelves its most dangerous model. Plus 10 more headlines, a new tool, and a security audit prompt.
Consortium européen pour les communs numériques: l'EDIC sur les rails www.zdnet.fr/blogs/l-espr... Le Français Laurent Rojey a été choisi comme directeur de l'EDIC. Chez les cinq Etats fondateurs, l’Agence de technologie souveraine de l'Allemagne recrute pour des projets open source.
@ritaraley.bsky.social and I wrote in cautionary terms (to say the least) about NotebookLM here. www.chronicle.com/article/ai-m...
Oracle licencie jusqu’à 30 000 personnes par mail pour financer son pari IA à 500 milliards
siecledigital.fr/2026/04/02/o...
2 Weeks of Iran War Released More Carbon Emissions Than 84 Countries Do Yearly truthout.org/articles/2-w...
Strikes on Iran are unleashing toxic pollutants that likely will linger for decades, poisoning the environment and posing health risks
www.bloomberg.com/news/feature...
🚨 BREAKING: German Chancellor Friedrich Merz has criticized the Trump administration for temporarily lifting sanctions on Russian oil in an effort to bring down surging energy prices amid the U.S. and Israeli attacks on Iran.
🔗 politico.eu/article/germ...
Why is the biggest name in AI late to the coding revolution?
@mzeff.bsky.social spoke to more than 30 people, including OpenAI execs and employees who spoke without company permission, about how OpenAI is racing to catch up with Anthropic -- and why it fell behind in the first place:
Nick Clegg, former president of global affairs at Meta, has a lot to say about tech under the Trump admin ("endless ring-kissing"), the EU's AI regulation ("a ludicrous act of self-harm"), and much more.
An excellent @wired.com interview from @joelkhalili.bsky.social.
. @stephanehorel.bsky.social qui enquête sur les PFAS a subi plusieurs tentatives de vols et a été suivi par des gens qui voulaient lui voler ses affaires de boulot.
ce que contient la plainte d’une journaliste du « Monde » après une série de vols
www.mediapart.fr/journal/fran...
Adieu l’installation d’APK : Google veut transformer Android en une prison dorée à la Apple www.frandroid.com/android/3008...
Le WSJ s'extasie dans un reportage photo sur une start-up IA créée par des ados (aujourd'hui 20, 19 et 17 ans) et valorisée 1 milliard de dollars.
Je me dis "bon, c'est qui les parents", je n'ai pas été déçue www.wsj.com/business/ai-...
Marre de Microsoft 365 et Google Workspace?
Cette nouvelle suite bureautique européenne 🇪🇺 est une alternative privée et open source.
#Office. eu , une alternative solide
buff.ly/sugDpAm
Palantir’s Double Conflict of Interest in the War Against Iran – Byline Times bylinetimes.com/2026/03/05/p...
www.youtube.com/watch?v=EMRJ...
youtube.com/shorts/6KOZO...
Apple’s new MacBook Neo gives the company its cheapest laptop ever, creating a budget entry into the Mac lineup.
TurboSparse Inference: 4.6x Faster LLM Decoding via Hybrid GPU-CPU Computing
Accelerate LLM inference with TurboSparse. Achieve up to 2.28x speedup on pure CPU and 4.64x in hybrid GPU-CPU environments compared to llama.cpp baselines.
#hackernews #llama #llm
I'm quietly rather delighted by how many people have responded to this one so positively across various places. The images do not show up in the preview block, so here is one of them
reasoning LLMs are honest about what they can do really well: deterministic inference, logical chains, consistent application of rules. but we're expecting them to be honest about something they fundamentally can't: human-like uncertainty in probabilistic judgments 🧐
Et si cette vidéo est proposée par les algos (youtube ...) maintenant, posez-vous des questions !
Cela date de 2024, c'est probablement une des sources utilisées par anthropic et c'est une perle en matière de prédiction et de stratégie :
www.youtube.com/watch?v=7y_h...
Elgato’s new Stream Deck Plus XL has 36 LCD buttons instead of 8, and 6 customizable dials. That’s a whole lot of Stream Deck www.theverge.com/tech/887761/...
split visualization: optimistic predictions (flying cars, AI) with unacknowledged constraints on left; successful predictions with considered constraints on right
why do technology predictions fail? not because they misjudge capability. because they're dishonest about constraints. flying cars are technically possible but regulatory, human behavioral, & economic constraints make them unviable. predictions that acknowledge constraints actually come true 🔮
Google va modifier ses résultats en Europe pour laisser plus de place à la concurrence.
Dans le détail, Google va lister non seulement ses propres résultats, mais aussi ceux des VSS ; les services de recherche verticaux les mieux classés seront proposés par défaut, selon une source de Reuters