NotebookLM experimentje gedaan, toespraak van WIlhelmina voor Radio Oranje ( 28-11-1944) geupload (geluidsfile, 197 MB). Na 2 minuten analyse van de toespraak, beschikbaar in Notebook, met een samenvattig etc, als ik er naar vraag ook een letterlijke transcriptie (!). De potentie van de deze tool voor analyse van content, ook voor audio, is enorm. En dan heb ik het nog niet eens over de veelgenoemde podcast functionaliteit. #AI #NotebookLM #audio

It's clear that Google has scored a hit with NotebookLM, 80.000+ organisations use it already, no longer experimental and new features are added. Personally i am also enthusiastic, best AI application since ChatGPT in my view. It also illustrates that having the most powerful model is not necessarily the only thing that counts.

Google

New in NotebookLM: Customizing your Audio Overviews and introducing NotebookLM Business

NotebookLM is piloting a way for teams to collaborate, and introducing a new way to customize Audio Overviews.

#AI #NotebookLM #GoogleGemini

Quantum technology is getting better in hacking encryption it seems, difficult to judge however from the outside how practical this all is.

The Quantum Insider

Chinese Scientists Report Using Quantum Computer to Hack Military-grade Encryption

Chinese researchers, using a D-Wave device, claim to have executed the first successful quantum attack on widely used encryption algorithms.

#quantum #encryption #military

I don't understand the obsession with the fact that LLMs/ AI models are not truly intelligent , don't understand anything and can not really reason (at least for now). That doesn't diminish their added value in many scenarions and applications. It's good to know the limitations described in the attached paper and take them into account, start working on mitigating these vulnerabilities. That will happen, we are not at the end of AI development. https://arxiv.org/pdf/2410.05229 #AI #reasoning

I use https://www.perplexity.ai/ for almost all my searches instead of Google, the way it presents it's sources makes it easy to check the information and the output is better then with Google. Recommended. #Perplexity #AI #search #google

Good sceptical article by Gary Marcus, regardless the fundamental problems with current LLMs , progress is remarkable however in my opinion. So reliable agents with LLMs are a problem, many other applications are not,

LLMs don’t do formal reasoning - and that is a HUGE problem

Important new study from Apple

#ai #garymarcus

Nu de Nobel prijzen bekend zijn kunnen we de discussie dat AI alleen maar een hype is, generatieve AI alleen maar een domme “ autocomplete” etcetera wel stoppen lijkt me. Los van het feit dat in de ontwikkeling van AI ook veel gebakken lucht zit maar dat kent elke technologie wel. #ai #nobel

Boeiend, ik krijg deze melding bij ChatGPT, het werkt overigens nog niet maar wellicht moet ik geduld hebben "Download de nieuwste iOS- of Android-app om de geavanceerde spraakmodus te proberen Voer natuurlijkere gesprekken in realtime met geavanceerde spraak. Voelt humor, sarcasme en meer aan en reageert erop." #AI #openAI #chatgpt #advancedvoicemode

Fascinating blog from May 2024 by Avital Balwit, working at Anthropic. Many observations i recognise, like this one, ""The general reaction to language models among knowledge workers is one of denial. They grasp at the ever diminishing number of places where such models still struggle, rather than noticing the ever-growing range of tasks where they have reached or passed human level" Also we are a bit in denial about the future impact of AI on our society in my view.

Palladium Magazine

My Last Five Years of Work

#AI

An excellent exampe of the amazing NotebookLM capabilities, this podcast about the Atlantic article, "I’m Running Out of Ways to Explain How Bad This Is" by Charlie Warzel. https://notebooklm.google.com/notebook/0630b38e-887e-49a9-9161-51bf76ea68a3/audio #notebookLM #AI #podcast

For all AI nerds out there like me, the yearly slidedeck, State of AI , always fun to browse through. #AI #StateofAI2024

Google Docs

State of AI Report - 2024 ONLINE

STATE OF AI REPORT. October 10, 2024 Nathan Benaich AIR STREET CAPITAL. airstreet.com stateof.ai

I am wondering about actual implementations and use of large generative AI models like ChatGPT, Gemini, Claude , Llama, Mistral etc in companies/organisations and how risks with regard to security, privacy, IP etc. are mitigated or controlled. Not a lot of case-studies are available it seems ? #AI #ChatGPT #Claude #Llama #implementation #security #privacy #IP

The nobel price for physics go to Hinton and Hopfield for their work in what we now call AI. https://www.nobelprize.org/uploads/2024/09/advanced-physicsprize2024.pdf #AI #ANN #neuralnetworks #hinton #hopfield #nobel

"I write about AI for a living — and NotebookLM is the most exciting tech to arrive since ChatGPT" , I agree...

Tom

I write about AI for a living — and NotebookLM is the most exciting tech to arrive since ChatGPT

The future of search is here!

#ai #notebooklm

A common mistake is that people think it's just about how powerful the models are in AI, while userinterface and way of working is just as important. That's the reason why NotebookLM and Open AI Canvas are leaps forward with regard to AI. #AI #OpenAI #Canvas #ChatGPT #NotebookLM

Good demo of the current risks with regard to AI / social media

The Verge

College students used Meta’s smart glasses to dox people in real time

Privacy will always be a major concern with smart glasses.

#ai #privacy #doxxing

Nog een preprint maar toch, Nederlands (!) onderzoek bij TU-Delft wat iedere gebruiker er van al wist, OpenAI o1-model is heel erg goed (bijna perfect) in Wiskunde. "It scored a near-perfect 76 and 74 out of 76 points. For context, only 24 out of 16,414 students in the Netherlands achieved a perfect score. By comparison, the GPT-4o model scored 66 and 62 out of 76, well above the Dutch average of 40.63 points." https://www.researchgate.net/publication/384071542_System_2_thinking_in_OpenAI's_o1-preview_model_Near-perfect_performance_on_a_mathematics_exam #AI #openai #GPT4o1 #TUdelft

Example of a NotebookLM podcast about two documents: "Hype, Sustainability, and the Price of the Bigger-is-Better Paradigm in AI" and "The intelligence age" by Sam Altman. It's nice how NotebookLM detects the contradictions and tensions, it starts talking about "balance", finding a middle ground, more attention to smaller models etc. Not perfect but good demonstration of it's capabilities. https://notebooklm.google.com/notebook/727d8ff3-56a0-4c87-a126-7e87db2f7805/audio #AI #NotebookLM #podcast

What if you want to organise a funding round with an extreme high target and want all investors to jump in because of FOMO ? You say "superintelligent AI could emerge in just a few thousand days, marking a potentially transformative moment in human history that could usher in an era of unprecedented prosperity and capability." 😀 #AI #SamAltman #intelligence #openai

Great idea. Submit Your Toughest Questions for Humanity's Last Exam.

Center for AI Safety

Submit Your Toughest Questions for Humanity's Last Exam | CAIS

CAIS and Scale AI are excited to announce the launch of Humanity's Last Exam, a project aimed at measuring how close we are to achieving expert-lev...

#ai #benchmarks