AI chatbots, who doesn't hate them 😆 , seem to have potential (giftlink)

When Your Building Super Is an A.I. Bot (Published 2024)

Artificial intelligence is doing everything from helping landlords communicate with tenants to managing energy use.

#AI #chatbots #NYT

How companies change their Terms of Service for AI. (giftlink)

When the Terms of Service Change to Make Way for A.I. Training (Published 2024)

Tech companies have been making subtle and not-so-subtle changes to their rules for better access to data for building A.I. We took a look at some ...

#AI #termsofservice #NYT #bigtech #privacy #IP

Good article about the datacenters/computing behind AI , for people interested in technology but even if you are not it gives a good view of the mind boggling scale with regard to energy, compute, memory etcetera https://www.semianalysis.com/p/100000-h100-clusters-power-network #AI #datacenter #compute

LLMs are amazing technology but i think they should not be used for directly delivering services to end-customers, for example in customer service. Many "accidents" already happened. They are simply too unreliable and hallucinatorily for that. The way ahead in my opinion lies in combination with humans and other tools/technologies/algorithms. But please don't connect your GPT4 chatbot with RAG directly to your customers 😆 #AI #LLM #chatbots

Consultants and AI, with some stories about actual implementations with ING and Dun & Bradstreet. (giftlink)

The A.I. Boom Has an Unlikely Early Winner: Wonky Consultants (Published 2024)

Rattled by tech’s latest trend, businesses have turned to advisers at Boston Consulting Group, McKinsey and KPMG for guidance on adopting generat...

#AI #consultants #NYT

So OpenAI played us, just like Google and others often do with regard to demonstrations, Voice Mode will be not generally available until fall this year. Not really surprising if you think about the compute that is probably needed in the background, still the continous overpromising will hurt OpenAI. #OpenAI #VoiceMode #AI

Interesting to see that besides general unreliability and that output can 't be trusted, different LLMs are best for different things. Claude 3.5 for coding, Perplexity for search and ChatGPT for many other things. The nice thing about the added video is that it reminded me about how much LLM's make up/hallucinate. Even complete URLs if you ask for sources within ChatGPT.... #AI #Claude #ChatGPT #Perplexity

A russian troll-farm ran out of ChatGPT credit, which resulted in this message on X. #disinformation #AI #Russia

Interesting start-up with Dutch co-founder.

TechCrunch

CuspAI raises $30M to create a GenAI-driven search engine for new materials | TechCrunch

The company says its platform functions like a search engine for materials, enabling the fast evaluation of a “vast number of novel structures.”

#ai #CuspAI

China is dominating in science. USA and EU have a problem... #china #science #geopolitics #economist #nature #eu #usa @npub17dnl...a7ch

Gazprombank is apparently down and authorities in Russia are probably trying to avoid a bank run. #Russia #gazprombank #bankrun #geopolitics

Private Cloud Compute by Apple, giving it the benefit of the doubt for now, a lot of promises "When we launch Private Cloud Compute, we’ll take the extraordinary step of making software images of every production build of PCC publicly available for security research", but at the same time it does not seem to be Open Source, so how can third parties check it enough ? 🤔

Private Cloud Compute: A new frontier for AI privacy in the cloud - Apple Security Research

Private Cloud Compute: A new frontier for AI privacy in the cloud - Apple Security Research

Secure and private AI processing in the cloud poses a formidable new challenge. To support advanced features of Apple Intelligence with larger foun...

#AI #Apple #privatecloud #PCC #security

Nice AI Act explorer.

The AI Act Explorer | EU Artificial Intelligence Act

#AI #AIact #EU #legal #legislation

Artificial intelligence and radiologists in prostate cancer detection on MRI (PI-CAI): an international, paired, non-inferiority, confirmatory study. https://www.thelancet.com/journals/lanonc/article/PIIS1470-2045(24)00220-1/fulltext #AI #medical #radiology #cancer

Crazy if you think about it. Afd , extreme right party in second place , in Germany. The country that experienced the NSDAP, Hitler etcetera #exitpoll #afd #germany

The difference between Israel and Hamas is clear, one a democracy (with problems) and the other is a terrorist organisation. At the same time it means Israel's actions should maintain high standards with regard to innocent victims. They seem to be failing in that respect... #gaza #hamas #israel

Orban doing Putin's work: "EU ambassadors fail to decide on launch of accession talks with Ukraine due to Hungary's objection"

Ukrainska Pravda

EU ambassadors fail to decide on launch of accession talks with Ukraine due to Hungary's objection – media

Permanent Representatives of the European Union member states have again failed to make a decision on the European Commission's recommendation...

#ukraine #eu #hungary #orban #geopolitics @npub17dnl...a7ch

GPT-4o never gives such terrible answers as Google AI in search that is being piloted, many terrible examples are shared in social media. Do they use a very simple model or is google's implementation just awfully bad ? 🤔 #Google #AI #Search

Orban now starting to actively undermine NATO.

UNITED24 Media

Hungary 'Reconsidering' NATO Membership to Avoid Non-Alliance Operations

#Orban #NATO #Geopolitics #Hungary #Ukraine @npub17dnl...a7ch

Although i am very enthusiastic about the functional aspects with regard to Microsoft Recall, it's a potential security nightmare. We should not be fooled with promises about local encrypted data, that can be compromised, also through federated learning, data could still reach microsoft. Also probably part of the compute will be in the cloud. #Microsoft #Recall #cybersecurity #privacy #AI