Good tips from the FBI, "Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud"
https://www.ic3.gov/PSA/2024/PSA241203
#cybersecurity #fraud #ai #financial
Confirms my own observations, GenAI is already everywhere. “Google poll says pretty much all of Gen Z is using AI for work. Some 82% of young adult knowledge workers said they leverage AI in their work, a new Google Workspace survey found.”
https://qz.com/google-workspace-poll-ai-workplace-gen-z-millennials-1851707141
#ai #workplace #genz
So a large part of content on Linkedin is AI generated, not surprising but the percentage is high (50%+).
https://www.theregister.com/2024/11/28/linkedin_ai_posts/
#AI #Linkedin
Great explanation of the opensource Model Context Protocol (MCP) made by Anthropic which could have a big impact. It connects your LLMs to other applications and datasources (bi-directional), which gives enormous possibilities. The good news is that everybody can start developing with it because it's opensource.
#Anthropic #Claude #MCP #AI #LLM
actually an interesting idea and proposal, MCP, model context protocol.
https://www.anthropic.com/news/model-context-protocol
#openstandard #opensource #mcp #ai #anthropic
Discovered I can play my Ubisoft games I purchased while I had #GoogleStadia , I can now play them for free on Amazon Luna because I have a Prime account and it's able to connect to Ubisoft. Nice to have AC Odyssey and AC Origins back , great games.
Still miss Stadia, what an easy and great service that was.
#cloudgaming #google #ubisoft #prime #games #stadia
Good example how AI can improve without necessarily improve the current models. Combining models, let them cooperate on a task could bring big improvements. For that a language is usefull and this preprint does a proposal for that.
#AI #Agents #Droidspeak
Genoeg te bespreken in de Ministerraad, bezoek Wilders aan Netanyahu en onderstaande, https://www.trouw.nl/binnenland/kabinet-ontkende-het-bestaan-van-asielcrisis-notitie-maar-deze-blijkt-er-wel-te-zijn~bc90f4a6/
#politiek #kabinet #schoof
Benieuwd of kamerleden inmiddels ook NotebookLM (of andere AI tools) gebruiken om de bergen aan documenten die zij ontvangen te analyseren.
Zie bijvoorbeeld het bizarre aantal moties van de 21e (via nostr:nprofile1qy2hwumn8ghj7un9d3shjtnddaehgu3wwp6kyqpqw2uuz5qqt22rguwke89qhzzuk0v0jcgs55z9h02a7hf58qr4lmjscn49de van nostr:nprofile1qy2hwumn8ghj7un9d3shjtnddaehgu3wwp6kyqpqjngvk9z4jh3awzcnc8m2duxf5knjwxxap7x78gypfsphlck3ff2q0a4az6 )
#TweedeKamer #AI #politiek
An interesting paper, AI helps scientists discover 44% more materials, leading to a 39% increase in patent filings and a 17% rise in product prototypes. The technology improves research efficiency by 13-15%, but its benefits are greater for more skilled scientists. According to this paper.
https://aidantr.github.io/files/AI_innovation.pdf
#AI #Science #innovation #materials
There are bottlenecks, but It's not that improvement has stopped on training current AI models, if you read papers like this, "The Surprising Effectiveness of Test-Time Training for Abstract Reasoning"
#AI #LLM #reasoning
The disruption and impact of generative AI will not be diminished by the fact that progress is slowing down with regard to the large AI models as widely reported. Current models will get cheaper, better and will be widely implemented, in many applications we already use.
#ai #generativeAI #impact
Very interesting,
Flexible and powerful framework for managing multiple AI agents and handling complex conversations,
#ai #aws #amazon #agents #llm
In a way it's great news that improving AI models by just scaling up seems to have run into problems. That will increase creativity and innovation to find new ways to improve models or maybe try new approaches. Also new combinations of technology can bring a lot.
#AI #scaling #innovation #ICT
I am using both GPT-4o and NotebookLM, i have the impression that NotebookLM is much better at summarizing and analyzing texts, i am wondering whether that's because of NotebookLM implementation aspects or because of that Gemini Pro 1.5 is better at this task then GPT-4o ? 🤔
#AI #NotebookLM #Geminipro #ChatGPT #openai #google #text #summarizing #analysing
Interesting, a german company, DeepL, launches DeepL Voice, real-time, text-based translations from voices and videos.
#AI #voice #text #translation #realtime
My workflow with very long youtube videos i want to check, is making a transcript with Tactiq.io , add the transcript to a notebook in NotebookLM , read the overview documents and ask additional questions if necessary, works like a charm for those long videos.
#notebookLM #youtube #AI
Nothing to get worried about, the top models at their current state are so good that they already can achieve great things. Especially when combined with other technologies and implementation techniques that limit their flaws. Also ofcourse literally billions are spend trying to prevent or circumvent limits for improvement.
#ai #scaling
https://arstechnica.com/ai/2024/11/what-if-ai-doesnt-just-keep-getting-better-forever/
nostr:npub146tl0yp5v2hgmz8rxcmkvwp6sn3jw70gg9p4zeyyg6hyvqtzcusqlqcfpw Bluesky heeft veel aantrekkelijke kanten ( @erikjonker.bsky.social ) maar Mastodon vind ik zelf toch fijner werken, ook is volstrekt onduidelijk hoe het verder met Bluesky gaat qua bedrijsmodel en je "vertrekt"/wisselt niet zo eenvoudig als Mastodon (je bent zo naar een andere instance vertrokken met al je volgers).