Nathaniel Manning’s Post

"To move forward, the debate must be recast, making it clear that protecting creative industries will help sustain the flow of quality information into the AI models of the future. Otherwise, the industry faces the scenario Siri creator Gruber described, known as 'model collapse' — where AI models train solely on AI-created materials, slowly degrading over time." https://lnkd.in/eRmRVhhq

Whether you're sending logs, metrics, traces, or events to Dynatrace, you can do so much more than put this data on a dashboard! This tutorial will show you how to spot traffic spikes and loads using seasonal and auto-adaptive baselining. You'll learn to forecast future trends, which will help you anticipate when your system will hit its limits or achieve your business goals. #observability #ai #davisai

🥁 Can AI Make Music Production Faster? HiQ joined Stockholm MLOps to explore how machine learning can transform the creative process in music production. Arthur Öhman, MSc student in Machine Learning at KTH Royal Institute of Technology and trainee at HiQ, presents a practical and research-driven deep dive into automatic drum transcription — turning raw audio into musical notation. In collaboration with Excellent Audio, Arthur developed a hybrid, interpretable, and efficient model combining Non-Negative Matrix Factorization (NMF), sigmoidal deconvolution, and CNN-based classification to enable a drag-and-drop workflow where musicians can convert drum audio to MIDI in real-time. He also discusses the motivation, model architecture, and trade-offs between speed and accuracy, along with broader applications in signal processing, EEG/MRI analysis, and anomaly detection — showing how ML bridges research and creativity. 🎥 Watch the full talk here: 👉 https://lnkd.in/eJTHyEaF This talk was recorded by our partner AI Sweden 🎬 — helping us share the incredible work happening in Sweden’s MLOps community. #StockholmMLOps #AI #MachineLearning #MusicTech #CreativeAI #SignalProcessing #Community #HiQ #KTH

For a Bloomberg Opinion bonanza series launching today, we set out to tackle some of the biggest global questions about AI. One of the thorniest is actually many difficult questions in one: ❓ Should AI companies be forced to pay for the copyrighted content they use to train their models? ❓ Should they be paid when that content is used to generate new output? ❓ How would that even work? Who handles the money? And the biggest: ❓ How do you stop it grinding AI innovation to a halt? My view on this aligns with Björn Ulvaeus, ABBA's legendary songwriter, who is an AI enthusiast but a leading advocate for creators getting what they're owed. "If you make money on something that I helped you create," he told me in a recent interview, "I get a share.” As I write in this essay, the US has the opportunity to set the global copyright agenda and protect the creative industries for generations to come. It's an opportunity it should take -- now. #ai #artificialintelligence #technology https://lnkd.in/eTNukjWJ

Looks like even browsers are trying to get their own slice of the AI pie—what's next, AI-powered snack menus? Opera is launching its AI browser, Neon, at $19.90 a month with limited access and a waitlist for others. The feature aims to blend cutting-edge AI with everyday browsing, offering early users a sneak peek into the future of internet navigation. It's a move that reflects the broader trend of integrating AI into user experience, much like how every company's trying to stay relevant amid rapid tech shifts. For product managers, this underscores the importance of balancing innovation with user onboarding—getting people excited before they get lost in the waitlist maze. Thanks to Elissa Welle for her insightful write-up on this tech leap! #AI #Browsers #Innovation #UserExperience First published: October 2025

Will huge AI companies with mega-bucks stifle human creativity and deprive those in the creative fields of their livelihood due to AI companies' mass theft of IP off the Internet? "Time is of the essence. AI models that pilfer the work of human thinkers and creators without payment — then offer the public an ersatz version of what they create — undermine their livelihoods. At a moment when industries like Hollywood and publishing are already struggling, some worry it could be a fatal blow. Ironically, that’s also a threat to the future of artificial intelligence itself. Failing to resolve the copyright conundrum in a way that works for all involved could be a “real problem” for innovation, warns Tom Gruber, inventor of the Siri AI assistant. “You’ll have nothing authentic to train on anymore.” The Creative Cost AI’s champions speak of its development in the same determined tones as the race to the moon or the building of the atom bomb... ...Another criterion of the fair use test is that the market for a copyright holder’s work is not harmed. Some industries say this harm is already being done. News outlets say Google’s AI Overviews, which synthesize content into a single answer within Google Search, have caused a plunge in readership. Fewer clicks equals fewer ad dollars, which are still the lifeblood of many publications. Penske Media Corp., publisher of Rolling Stone, is suing, saying AI Overviews has “profoundly harmful effects on the overall quality and quantity of the information accessible on the internet.” Hollywood’s talent pipeline is also suffering. Sam Tung is an LA-based storyboard artist whose clients have included Marvel, A24, Nintendo and Warner Bros. He also co-chairs the AI Committee of the Animation Guild, an organization born in the aftermath of Walt Disney’s efforts in the 1950s to ship some of his cartoonists’ labor overseas. Members worry that entry-level gigs are disappearing, says Tung. At the same time, established professionals are reporting demands that projects be kept shorter and cheaper, as more clients suggest the work can be streamlined using AI." #AI #copyright #theft #fairuse https://lnkd.in/eCmEswXn?

🎵 Turning earworms into learning tools. In AEE 216: Practical Parliamentary Procedure, memorizing 25 motions (and their precedence, debatability, vote thresholds, etc.) can overwhelm students. So we tried something different: using GenAI (Suno + the Penn State University Teaching and Learning with Technology AI Arcade) to convert key concepts into songs. My TAs, Alexis Otstott and Mary Ribbeck, built a five-track “playlist”: one for Main Motion and one for each class of secondary motions—each in a different genre to make recall stick. Why this matters: 🎯 Retrieval cues > rote lists 🧠 Dual coding (music + text) 🧩 Chunking by motion class Take a listen and tell us what hits, what misses, and what to try next. How else can we utilize AI to make parliamentary learning more engaging and enjoyable? 👉 https://lnkd.in/eENrhS9k #ParliamentaryProcedure #TeachAg #TeachingWithAI #LearningScience #HigherEd #EdTech #AIinEducation #PennState Penn State College of Agricultural Sciences

SoundPatrol has launched an AI-enabled anti-piracy research lab. Foley & Lardner LLP represented SoundPatrol in the formation and financing of this lab, which aims to protect music for artists and labels. The lab, co-founded by Walter DE BROUWER and Michael Ovitz, focuses on large music models and neural fingerprinting to combat music piracy. SoundPatrol has partnered with Universal Music Group and Sony Music Entertainment. Read more: https://lnkd.in/ebEmBNjT 📰 Subscribe to the Daily AI Brief: https://lnkd.in/eBqX6ru5 #ai #artificialintelligence #ainews #LegalAI

Responsible AI Associations: #AIforMusic - The Principles for Music Creation with AI establish a guideline for the responsible use of #ArtificialIntelligence in music. By adopting these principles, companies and organizations commit to protecting the vital contributions of human creativity and to the responsible development and application of artificial intelligence for #musiccreation. Scroll down for the list of supporters. https://aiformusic.info/ #ai4music