Avatar
Erik Jonker
ddc6c81c03da216550654f73121985d8f30636aac98903de01993746bab7bdb3
Civil servant from the Netherlands, fascinated by innovative technologies and how they can improve society. Also interested in geopolitics and international affairs. All opinions expressed are purely personal and do not represent the opinion of my employer or anyone else. #digitalgovernment #innovation #digitaltransformation #AI #technology #tfr tootfinder #ICT #IT #fedi22 #geopolitics #AcICT Signal: erikjonker.09

Nice work by MIT and a good overview/start for thinking about risks with regard to AI. The AI Risk Database captures 700+ risks extracted from 43 existing frameworks, with quotes and page numbers.

https://airisk.mit.edu/

#AI #AIrisk #repository #MIT

I am just wondering if Gemini Live is already available (as promised) for anyone here with a paid Gemini advanced subscription?

I am thinking about trying it for a month but not if it's not available yet 😃

#AI #GoogleGemini #GeminiLive

Experimental but fascinating, "| The AI Scientist: Towards Fully Automated

Open-Ended Scientific Discovery"

https://github.com/SakanaAI/AI-Scientist

#AI #SakanaAI

I am impressed with STORM, an AI demo by Stanford that generates articles about subjects you want. You can login with a google account and try it out.

https://storm.genie.stanford.edu/

#AI #Stanford #STORM

Nuttige oproep van Autoriteit Persoonsgegevens met betrekking tot gebruik van ChatBots.

https://autoriteitpersoonsgegevens.nl/actueel/let-op-gebruik-ai-chatbot-kan-leiden-tot-datalekken

#AI #AP #Chatbots #datalekken #privacy

Scary, imagine what is possible in fraud and criminal activities with this now broadly available technology. It's simple, you can't trust a face in a videoconference.

https://www.geekersdigest.com/deep-live-cam-live-swaps-your-face-with-another/

#AI #crime #fraud #deeplive #cybercrime

Interesting new problems on Spotify with regard to AI generated music.

https://www.fastcompany.com/91170296/spotify-ai-music

#ai #spotify #music

Nice overview of AI Chatbots. Don't focus too much on benchmarks, different Chatbots are good in different things and in different contexts.

#ai #chatbots

https://zapier.com/blog/best-ai-chatbot/

ChatGPT standaard in diverse Peugeots (wel tegen betaling per maand), vraag me af welke grip iedereen heeft op data verzameling en gebruik bij dit soort toepassingen. 🤔

https://www.media.stellantis.com/nl-nl/peugeot/press/peugeot-integreert-chat-gpt-standaard-in-complete-modellengamma

#ChatGPT #Peugeot #AI

Interesting, GitHub - stanford-oval/storm: An LLM-powered knowledge curation system that researches a topic and generates a full-length report with citations.

https://github.com/stanford-oval/storm?tab=readme-ov-file

#ai #storm #llm

AI algorithms are amazing but this article makes you think about vulnerabilities that can't be fixed.

https://arstechnica.com/ai/2024/07/superhuman-go-ais-still-have-trouble-defending-against-these-simple-exploits/

#AI #vulnerabilities

Even if LLMs wouldn't get any better (and they will), still development of AI applications will continue at high speed because of increasing efficiency in various algorithms and processes in the background. This will make more applications possible and affordable. Minference by Microsoft is a good example.

#AI #Microsoft #Minference

Nice article in Nature, "Can AI be superhuman? Flaws in top gaming bot cast doubt" , extrapolate the mentioned flaws to current implementations of AI and you can see all the risks we are facing.

https://www.nature.com/articles/d41586-024-02218-7

#AI #risks #Nature

Nice paper,

"No Need to Lift a Finger Anymore? Assessing the Quality of Code Generation by ChatGPT"

https://arxiv.org/abs/2308.04838

#AI #ChatGPT #code #coding #programming #LLM

Interesting article about the actual added value of generative AI for intelligence agencies.

"Raman said U.S. intelligence analysts currently use generative AI in classified settings for search and discovery assistance, writing assistance, ideation, brainstorming and helping generate counter arguments."

https://www.defenseone.com/technology/2024/07/us-intelligence-community-embracing-generative-ai/397860/

#AI #intelligence

https://websim.ai/ is not necessarily useful for anything but it's fun and interesting, also it shows what AI already can do and made me think about the future.

(Also it's free 😀)

#ai #websimai

I can fully agree, also my experience: Google has been overpromising — and under-delivering — with Gemini.

https://techcrunch.com/2024/06/29/geminis-data-analyzing-abilities-arent-as-good-as-google-claims/

#AI #Google #Gemini