techhub.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A hub primarily for passionate technologists, but everyone is welcome

Administered by:

Server stats:

5.4K
active users

#generativeAI

243 posts127 participants6 posts today
Continued thread

⚙️ Quindi, il funzionamento l'abbiamo intuito, e possiamo anche pensare di usare modelli per misurare la pertinenza dei contenuti alle query per ragionare sul "posizionamento" su AI Overviews.
📈 Chiaramente, il primo step rimane l'essere tra i risultati rilevanti, che rappresentano la knowledge a disposizione del RAG.

___ 

✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: bit.ly/newsletter-alessiopomar 

Alessio PomaroAlessio Pomaro, Head of AI, Docente, Speaker, AutoreAlessio Pomaro, Head of AI, Docente, Speaker. Ingegnere, LinkedIn Top Voice Italy e autore di Brand Voice (FrancoAngeli Editore) e Voice Technology (Dario Flaccovio Editore).

“The story of automation in the US is that it has mostly impacted on manual workers in manufacturing. For example, factory employees — such as carmakers — performing routine tasks have lost their jobs to robots — or lower-cost Asian competitors.

#IndustrialAutomation has tended to affect lower-skilled, #BlueCollar jobs in the “#rustbelt” heartlands and small-town, less-educated communities in the south and midwest.

But a recent study from the #BrookingsInstitution suggests that the communities most exposed to AI-driven job dislocation will be #WhiteCollar information workers. The researchers studied the usage of #OpenAI’s #GenerativeAI tools across more than 1,000 occupations and mapped this against where those jobs were most commonly located.

Their analysis suggests that many #coders, #lawyers, #FinancialAnalysts and #bureaucrats in cities such as San Jose, San Francisco, Durham, New York and Washington DC might want to rethink their futures. But #NonOffice-bound #workers in places such as Las Vegas, Toledo, Ohio and Fort Wayne, Indiana may be less exposed to AI disruption.”

My observation since 2022 when #ChristopherHohn an influential shareholder decided to *speak out* about “reducing its head count and paying (hi-tech) workers less”. [1]

This is the decade where extreme (cost) pressure will be forced on White collar workers by the introduction of AI.

<archive.md/YqF03> / <ft.com/content/04343a69-8204-4> (paywall)

[1] <forbes.com/sites/jonathanponci>

📢 AI-Phi Session 22 – Beyond Consent in Generative AI 📢

Can we truly consent to AI-generated content when our data creates representations beyond our control? 🤖⚖️

Join Giada Pistilli as she explores the "consent gap" in AI, questioning traditional legal & ethical frameworks.

📅 April 17, 2025 | ⏰ 11 AM
📍 Sony CSL, Paris
🔗 Details & Registration: ai-phi.github.io/posts/session

#AI #Ethics #GenerativeAI #Philosophy #AI-Phi #Paris

ai-phi.github.io · Session 22 - Giada Pistilli - Beyond Consent, The Limits of Data Agency in Generative AI SystemsCan traditional notions of consent meaningfully apply to AI-generated content? What happens when our data creates representations of ourselves beyond our control?

If you’re a leader betting on AI, you’re not just managing tech—you’re managing trust.

A new Pew study shows a major gap:

🧠 79% of AI experts see a positive or neutral future with AI.
😟 But only 50% of U.S. adults agree.

That’s a trust deficit.

Customers want control, are wary of bias... and they’re watching.

If AI touches your customer experience, you’ll need clear communication, transparency, and genuine trust-building.

🔗 Full study:
pewresearch.org/internet/2025/

Pew Research Center · How the U.S. Public and AI Experts View Artificial IntelligenceBy Beshay

Day 4 of the 5-day Generative AI intensive course by Kaggle and Google covered, among other things, Search Grounding:

:blobcoffee: With this technique, you can enrich generative AI models with up-to-date and reliable information from the internet — in this case by connecting them to Google Search.

That means the model no longer relies solely on its training data but also integrates information retrieved from search results.

:blobcoffee: So what's the difference to RAG (Retrieval-Augmented Generation)?

There’s no need to build and manage your own retrieval system (like a vector database), since the additional data is fetched via an API from a search engine.

Here is the link to the corresponding Kaggle notebook: kaggle.com/code/markishere/day

www.kaggle.comDay 4 - Google Search groundingExplore and run machine learning code with Kaggle Notebooks | Using data from No attached data sources
#ai#GenerativeAI#ki

Hard to see Zucherberg's Meta improving it's already low image, or that of Generative AI, by this move. They just don't care, this is deliberate theft to help recoup the $$$Bn on building capability greater than any realistic revenue model.

Another good reason to turn off training permissions in any AI tool use with family history documents and source mages. But I still wouldn't trust Meta.

#AI #generativeAI #CopyrightTheft #LiteraryTheft #genealogy ##familyhistory

theguardian.com/books/2025/apr

The Guardian · ‘Meta has stolen books’: authors to protest in London against AI trained using ‘shadow library’By Ella Creamer

🧠 #OpenAI ha lanciato ufficialmente l'#AI Academy: una piattaforma formativa gratuita pensata per chi vuole portare l’AI dal laboratorio alla realtà operativa.
✨ I contenuti sono pensati per l’uso pratico: automazioni, agenti AI, pipeline con GraphRAG, Q&A su documenti, integrazioni reali con diversi modelli.

🔗 L'Academy: academy.openai.com/home

___ 

✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: bit.ly/newsletter-alessiopomar