Quinton Krueger

A recent article in MIT Technology Review points out that the marginal gains from massive, cutting-edge models are starting to diminish. In response, smaller models—trained on more focused datasets—are now performing on par with, or even outperforming, their larger counterparts. https://lnkd.in/epdpJmHt

34

Explore #AI innovations in digital work activity capture with Vincent La! Learn about UTBMS Code Prediction and Bill Creation with Narrative Generation. Discover how #Airflow helps with model retraining, redeployment, and cost management. Register at: https://airflowsummit.org/ #ApacheAirflow #LegalTech

24

1 Comment

Working with the Palantir Technologies team, across multiple shared customers, has been a great experience for Physna. They do not partner in the usual way - here are three examples: 1. Speed: Palantir is a larger company, but they genuinely operate at series A/B speed. The way they build product and GTM is unique, as an operating model, and allows them to move unusually fast. 2. Scrappy, customer obsessed: Like early stage startups, they Obsess over customer problems. They invest upfront in new customers, and unlike most companies - they are support production engineers going onsite, to solve problems in non Palantir (eg. on prem, customer) environments. 3. Partnering approach: In our experience, the way they partner is mutually beneficial on day 1. Instead of waiting for you to bring them business, they go out of their way to make you successful on their platform (bringing you leads). Special thanks to Nikhil Khosla and Aaron Jaffe, who are both exceptional talents - and represent what makes Palantir a different type of partner. If you work at a startup focusing on Defense or F500 segments, consider if there is an opportunity to work with them. Our partnership has been great, and we are looking forward to expanding it:

36

2 Comments

This is an interesting idea, though I'm not sure LLMs are ready for a lead roll. Paper-bots, or AI agents for different aspects of science, have been kicking around since before the public release of ChatGPT a couple years ago (e.g. Microsoft Research's BioGPT effort from 2022). But the devil is in the details and the prospective referencing ability it would provide requires more than what the current generation of LLMs (or any likely extrapolation of these) can provide. Making up references relegates you to a supervised support role at best. There are better ways for such a specialized and highly sensitive use case. Hopefully these will lead us on the next stage to the #unjournaling. Otherwise we're going to have a lot more bots writing fan-fiction science. We need more signal, less noise in science. #science #papers #bots #agents #research https://lnkd.in/e8Jyrp3S

14

4 Comments

The research that Nicholas King highlights here shows how looking beyond the true scope of AI - it is much broader than many people are considering. While the hype is concentrated around Copilot, ChatGPT and Claude, many business use cases need a broader view. And Data Kinetic is here to help you build that view and find the Applied AI approaches and models that work for specific applications.

2

This state-of-the-art device could cut down AI's obscene energy use by a factor of at least one thousand. Here's how University of Minnesota engineers developed this hardware. #AI #ML #futurism #IntelligenceFactory #digitaltransformation #DX https://lnkd.in/ghZNXGBU

3

Join Carnegie Mellon University's Philip Koopman on Sept 11, 10 AM EDT, for insights on leveraging #AI effectively. Learn about: • Classification & generative outputs • Addressing bias & safety • The 90/10 principle • Real-world Machine Learning applications Understand AI's potential and limitations. Register now! https://lnkd.in/gvhVk33y #MachineLearning #TechInnovation #BOSS2024 #ArtificialIntelligence #DeepLearning #Semiconductor #ML

8

Check out this week's episode of #AdventuresinMachineLearning with Brooke Wenig #ML: Balancing Theoretical Knowledge with Hands-on Experience https://lnkd.in/gdmSCYap

2

Solutions Review’s “InsightJam.com” (US) interviewed Imran Chaudhri as part of a panel discussion about the dynamic intersection of AI and graph databases and their combined potential to revolutionize data management and application development. Watch Now: https://lnkd.in/gyHnT3RZ #datamanagement #dataplatform #databases #AI

8

📢 Are AI models like ChatGPT just “stochastic parrots”? This question takes center stage in a fascinating debate about the true nature of large language models (LLMs). Are these systems simply sophisticated pattern mimickers or do they hint at genuine understanding beneath the surface?   In his latest article, Andrew McAfee explores the critique that LLMs, despite their fluency, may lack real comprehension, producing text that mirrors their training data without deeper insight, raising doubts about their reliability.   In the other hand, experiments like Othello-GPT show that LLMs can form internal models of the world—such as understanding game rules—suggesting they might grasp more than just patterns. However, this emerging capability comes with ethical and legal challenges, emphasizing the need for careful consideration before deploying these systems at scale.   Can we really trust these systems to understand or is it all just a cleverly crafted illusion? 🤔   #ArtificialIntelligence #MachineLearning #AIModels #TechnologyInnovation

3

1 Comment

🚀 Generative AI: A Game-Changer for Strategic Insights 💡 Imagine if you could simulate complex scenarios, predict outcomes, and refine your strategies with pinpoint accuracy. Sounds like science fiction, right? Well, it’s not. Generative AI, like GPT-4, is here, and it’s transforming the way we approach strategic analysis. Think of it as wargaming on steroids – not just for military use anymore but for business, cybersecurity, and crisis management too. Why is this a big deal? Here’s the scoop: Speed: Generate and analyze scenarios in no time. Accessibility: These tools are now user-friendly and cost-effective. Even small businesses can get in on the action. Scalability: Tailor them to fit any industry or organization size. But with great power comes great responsibility. Ethical considerations are paramount – ensuring data privacy, fairness, and transparency is crucial. Ready to dive in? Here’s how: 1) Choose the Right Tools: Find AI tools that match your needs. 2)Train Your Team: Equip them with the know-how. 3) Pilot and Iterate: Start small, gather feedback, and refine. Generative AI isn’t just a tool; it’s a shift in how we think, plan, and act. Are you ready to embrace the future of strategic insights? Let’s do this. 🚀 🚦 #AI #ArtificialIntelligence #GenAI #AIInnovation #AIforbusiness #Strategy #wargaming #AItrends #scalable #Futureofwork

8

This week, Joe Gottlieb, Higher Digital's President and CTO, sits down with Miro Humer, Vice President and CIO at Case Western Reserve University, a top-tier research institution in the U.S. 🏛️ Miro shares how Case Western is leveraging objective prioritization to advance its renowned research agenda, foster trust and collaboration across campus, and embrace cutting-edge tools like the HoloLens and AI to revolutionize teaching and discovery. 💡 If you’re a higher ed leader looking to learn from a world-class research university’s approach to balancing IT resources, empowering innovation, and building a culture of collaboration, this episode is for you! #DigitalTransformation #HigherEducation #EdTech #ResearchUniversity 👉 Listen here https://lnkd.in/gzdnDbnG Tune in to TRANSFORMED wherever you get your podcasts and join the conversation on strategic IT leadership in higher education.

8

New interview with our CSO Dr. Matthew Ikle examines the current trajectory of AGI development, explores AGI timelines, contrasts neural-symbolic approaches with DNN architectures, and analyzes the convergence of AI and blockchain. Learn more: https://lnkd.in/ecHdJ3rr

21

This state-of-the-art device could cut down AI's obscene energy use by a factor of at least one thousand. Here's how University of Minnesota engineers developed this hardware. #AI #ML #futurism #IntelligenceFactory #digitaltransformation #DX https://lnkd.in/g5xRSDdX

4

Read 2023 #MostRead paper of TAAS, "Modeling, Replicating, and Predicting Human Behavior: A Survey." This article provides a succinct yet systematic review of important approaches in two areas dealing with quantitative models of human behaviors. https://bit.ly/3UzrFKR Andrew Fuchs, Andrea Passarella, Marco Conti #machinelearning #openaccess

3