Nicola De Coppi

AI is navigating a very tricky transition from "novel" to "useful" at the moment, and while the wild creative swings and the (now-maligned) wearables are capturing the attention, there's a deeper discussion happening about what it means to get generative models ready for reliable, robust, enterprise-grade deployment. More in our breakdown of this week's top #fashiontechnology news: interli.net/4aYFUhn

38

4 Comments

💡What strategies distinguish high-performing companies today? Great insights from analyzing pairs like NVIDIA vs. Intel Corporation, Microsoft vs. Oracle, Netflix vs. Paramount - you get the pattern.😉 1️⃣ Borrow someone’s road - leverage existing platforms to accelerate growth. 2️⃣ Partner with a third party - collaborate to expand capabilities and reach. 3️⃣ Reveal your strategy - transparency leads to trust. 4️⃣ Be good - ethical practices are a business imperative today. 5️⃣ Let the competition go - focus on your unique path, not on rivals. 6️⃣ Adopt small-scale attacks - incremental innovation leads to significant disruption. ❇In essence, collaboration and partnerships, ethical practices, and focus on innovation seem to be the winning formula for these companies. And given they have been outperforming their rivals on revenue growth, operating margin, and enterprise value growth for at least 5 years, there is definitely a lesson to be learned there. What is your favorite strategy for business success? Check out the Harvard Business Review's article by Kaihan Krippendorff: https://lnkd.in/e8ApUfWT #BusinessStrategy #Innovation #Leadership

13

2 Comments

🤔Some startups or SMBs still wonder if they should consider and implement generative AI now. It implies changes and costs that are not insignificant indeed. Yet, the question is not "if". It should be "how". 👏This case of a fintech leveraging generative AI is inspiring as it gives some answers to how this technology should become a true asset when a strategic AI readiness and effectiveness approach is taken 💡They automated document processing further and used data that were created during previous digital transformation efforts, specifically data coming from digitized documents 💡They focused on the refinement and prescription of prompts to increase generative AI effectiveness and, therefore, the level of AI-generated content accuracy and relevance 💡They targeted tasks where generative was expected to create most value associated with more or less human supervision/verification ➡Their focus on prompts is really interesting as it highlights a great way of maximing the power of LLMs. The following quote is spot on to demonstrate the value of language and domain expertise⬇ "Effective prompt engineering goes beyond simply crafting the perfect string of words. It involves: *Understanding the specific business context and requirements *Developing strategies to maintain AI system accuracy and reliability *Creating complex multi-step prompting strategies for advanced tasks like document processing *Collaborating with domain experts in finance and risk assessment to incorporate specialized knowledge into AI interactions"

15

3 Comments

New Creation... a small language model for Consulting https://lnkd.in/gVuqRQNP this is just a toy model to test the ability to create custom agents using the LLAMA3 70B model. This is an example of how open source is driving Innovation and speed... it really is a small exercise to create something simple. click the link and give it a try it's focused on Consulting based on a very simple instruction set and some very good Consultants who know how to explain things clearer than I do. you start with asking you to look at your company's website and give it a problem for to help solve. it will try and provide some instruction or some direction on Consulting techniques to solve that problem. for example... use the following prompt about a company called grow.com which is a pretty good company that does dashboards and Analytics. prompt: review my company website: https://grow.com assess the clarity of our messaging on the website and describe a summary of the business and our key customer segments along with any competitors we have in that segment. that should give you a good starting point.

3

This is a bigger threat to the future of science than any potential budget cuts. See UPDATE in comments (spoiler - it's worse than we thought). Bad business models around a flawed peer-review system plus wanton use of half-baked AI have left science in perilous situation. In the Dark Ages there were at least monasteries to hide and preserve science. With the current threat - the "shittification" of science to borrow Gary Marcus ' term - the future of science is in jeopardy as all of our pipeline of knowledge is being infected. It will take a new application of something like the Republic of Letters to get us past this threat, with tamper-resistant validation and authentication of data, information (manuscripts), and knowledge (systematic reviews) along with verified and confidence scored tools, devices, models, people and organizations. Trust - but verify - the science. Hopefully we can still save it from ourselves. #science #research #publishing #technology #threat #future Heather Leigh Flannery Yoram Vodovotz Thomas C. Collin, PhD Amicia D. Elliott, Ph.D. Kris Rockwell Keith Main Paul Tibbits, MD Sandy Barsky Ivor Dsouza

25

19 Comments

🧠 AI is revolutionizing ADHD treatment. Here's why this matters: AI assistants provide 24/7 support for organising thoughts Diagnostic accuracy reaches 96% with AI analysis Virtual reality + AI creates personalized treatment plans Real-time behavioral monitoring optimizes care Data-driven insights improve patient outcomes But here's the truth: The real power isn't in AI alone. It's in the partnership between human intuition and machine intelligence. Agree? Disagree? Share your thoughts below. 👇 #ADHD #ArtificialIntelligence #HealthTech #Neurodiversity

2

1 Comment

While most business leaders are striving to grasp an understanding of the potential impact of Generative AI for their companies and customers speakers at The Wall Street Journal event are already focused on what is coming next, which they call Artificial General Intelligence (#AGI). They define AGI as the state where AI can perform all human cognitive skills better than the smartest human. Learn about artificial generative intelligence here! https://lnkd.in/gyRtivxv #genAI #analytics

19

1 Comment

💡 A bit surprising to see so much focus on strategic alliances for #genai from this McKinsey & Company article. ⏩ But the speed of change is so fast and so dramatic that building alliances is the only mechanism that could potentially make it work. 🐫 🐢 Compared to the speeds, with which businesses are capable of reacting, coupled with still relatively low levels of awareness and understanding, the pressure of #aiadoption is rising. 😲 And, what we currently have are only hypotheses. No one knows for sure what will happen and how it will happen. #genai #aiadoption #futureofwork https://lnkd.in/guV7N7uW

4

5 Real Scenario to Prove “When data meets research, innovation begins" https://lnkd.in/geuyPWNM #DataInnovation #ResearchAndData #DataDriven #InnovationStartsWithData #RealScenarios #complereinfosystem

5

1 Comment

Executive Summary - Key Conclusions AI's Current and Future Capabilities: 🌟 AI technology, especially in the form of generative AI and large language models (LLMs), is rapidly advancing, with significant potential yet to be fully realized in commercial applications. The capabilities that exist today are not fully deployed, indicating a considerable opportunity for productivity gains in the near future​. 🌟 Andrew Ng emphasized that AI agents, or "agentic workflows," which involve iterative processes and error correction, are likely to be a key area of growth. These workflows allow AI to refine outputs over multiple iterations​. Open Source vs. Proprietary Models: 🌟 Ng highlighted the debate between open-source AI models and proprietary systems. He argued that open-source AI has significant advantages, including fostering innovation and making AI more accessible globally. He also pointed out that fears around the misuse of open-source AI (e.g., for creating bioweapons) have largely been exaggerated by certain lobbying groups​. 🌟 The economic viability of proprietary models is increasingly challenged by the rapid decline in training costs, which erodes the competitive advantage of large, expensive models. Implications for Business Strategy: 🌟 The evolution of AI will likely shift how businesses operate, particularly in knowledge-intensive industries. AI can handle more complex tasks, reduce the need for human intervention in decision-making processes, and transform workflows in sectors like healthcare, finance, and logistics​. 🌟 Companies need to consider how to integrate AI into their operations strategically, ensuring that they are not only adopting the latest technologies but also preparing their workforce and systems for the changes these technologies will bring​. Action Items Accelerate AI Adoption: 🌟 Identify key areas within the organization where AI, particularly agentic workflows, can be integrated to improve efficiency and decision-making. Invest in AI infrastructure. Explore Open-Source AI: 🌟 Consider the benefits of incorporating open-source AI models into the company's technology stack. Evaluate the potential for cost savings and innovation. 🌟 Stay informed about the latest developments in AI legislation and the broader industry debate to ensure that the company is positioned to adapt to changes in the regulatory environment. Enhance Workforce Training and Development: 🌟 Develop training programs to upskill employees in AI and machine learning, focusing on how to work with AI tools. 🌟 Encourage cross-functional collaboration between technical teams and other departments. Monitor and Evaluate AI Integration: 🌟 Establish metrics to assess the impact of AI on productivity, decision-making, and overall business performance. 🌟 Engage with stakeholders across the organization to gather feedback and insights on AI deployment. #learningdispatch #aieducation #ai #artificialintelligence #machinelearning #futureofwork

4

1 Comment

There are a lot of excellent RAG strategies in this blog post from Anthropic. There is not currently a true "best practice" or "golden" benchmark to retrieval mechanics . However, this covers a lot of components we at DiligentIQ employ (and have found successful) as part of our architecture for optimizing embedding/retrieval logic for understanding due diligence documents found in a VDR. The biggest challenge for many organizations is trying to overlay generative AI to the mountains of documents stored across an enterprise. This is a very hard problem to solve in 2024. Largely because document management is not something most firms do well and recency / relevancy is a complicated issue. My best recommendation is don't be so ambitious! Start with a finite set of documents where you can experiment, validate and explore the possibilities of this novel technology. Walk before you run. It's changing so fast, you're better off learning quickly and being nimble vs. trying to solve "global galactic" problems ... which is never fun and always hard.

9

I found the article "Do You Really Need a #ChiefAIOfficer?" (https://lnkd.in/dWY4D584) from MIT Sloan Management Review to be particularly insightful. The discussion on the role of a CAIO in navigating the complex and rapidly evolving AI landscape is timely and crucial. The article highlights several compelling benefits of appointing a CAIO: ✅ reducing AI #fragmentation, ✅ driving transformational #change, ✅ managing AI #risks centrally. ✅achieving scalable solutions across the organization. However, the article also outlines some challenges: ⭕potential cross-functional conflicts ⭕risk of overemphasizing AI as the go-to solution for every business problem. While these are valid concerns, I believe that with clear mandates, strong #communication, and a #strategicfocus on AI's business impact, the benefits far outweigh the drawbacks. Ultimately, the decision to appoint a CAIO should be closely aligned with the organization's AI maturity and strategic goals. 📢What are your thoughts on the need for a CAIO in today's AI-driven world? #AI #Leadership #DigitalTransformation #businessImpact

36

Microsoft has announced the launch of new adapted AI models tailored to meet the unique needs of various industries. #datascience #AI #artificialintelligence https://hubs.li/Q02ZmqF10 If you enjoyed this post, follow me for more insights on AI and data science!

12

Exploring Frontier AI Research The insights by Frontier AI Research reveal significant trends and developments in the training and computing costs of AI models. Here are some key highlights: - Training Costs Over Time: The estimated training costs of various AI models from 2016 to 2023 show a clear upward trajectory, highlighting the increasing investment required to develop advanced AI systems. Models like #GeminiUltra and #GPT4 are among the most costly, reflecting the demand for cutting-edge capabilities. - Foundation Models by Organization: In 2023, Google leads the pack with the highest number of foundation models, followed closely by Microsoft and OpenAI. This underscores the competitive landscape in AI research and development, where organizations are racing to create and refine their models. - Training Compute by Domain: The training compute for notable machine learning models across different domains (language, vision, and multimodal) indicates a shift towards more complex and capable AI systems. The data showcases how investments in training computing are expanding rapidly, particularly in language fashions. - Diverse Research Contributions: The distribution of foundation models across industry, academia, and nonprofits highlights the collaborative nature of AI research. Organizations like #Anthropic and #HuggingFace are pivotal in driving innovation alongside tech giants. These insights underscore the dynamic nature of AI research and the significant resources being allocated to push the boundaries of what's possible in artificial intelligence. Thank you Stanford Institute for Human-Centered Artificial Intelligence (HAI) Resource: https://lnkd.in/d8QbZXZt Zvonimir Filjak Tony Moroney Dr. Martha Boeckenfeld Sam Boboev Prasanna Lohar Monica Jasuja #AI #ML #machinelearning #datascience #frontierAI #innovation #techtrends

26

If you'd like to get a true grasp of the possibilities that #AI holds for the future, I highly recommend giving this post by Osvaldo Coto a read. The #TedTalk will amaze. #ML #DeepMinds

1

1 Comment

Recent research from the BCG Henderson Institute (BHI), conducted by François Candelon, Lisa Krayer, PhD, Saran Rajendran, and David Zuluaga Martínez, has sparked an important conversation about the use of AI in business problem-solving. While AI holds immense promise, their study revealed a surprising finding: individuals relying on GPT-4 actually performed 23% worse at tackling complex business problems compared to those who didn't use AI at all. 𝐊𝐞𝐲 𝐓𝐚𝐤𝐞𝐚𝐰𝐚𝐲: AI is a powerful tool, not a replacement for human judgment. While AI can augment our capabilities, it's essential to approach AI-generated solutions with a critical eye and leverage our own experience and expertise. The most innovative solutions often arise from a combination of human creativity and AI assistance. So, what does this mean for businesses seeking to leverage AI effectively? It means finding the right balance between AI-driven insights and human expertise. This is where BarRaiser Interview Intelligence comes in. BarRaiser is a prime example of how to effectively leverage AI in the business world. By empowering interviewers with AI-driven insights, fostering collaboration, and making interviews more robust, BarRaiser is revolutionizing the interview process and helping companies make better hiring decisions. AI has the potential to transform the way we do business, but it's crucial to use it wisely. By leveraging AI as a tool to enhance human capabilities, we can achieve greater success and drive innovation. BarRaiser exemplifies this approach by redefining the balance between AI and interviewer expertise in the interview process. By approaching AI with a balanced perspective and using it strategically, businesses can unlock its full potential and achieve new levels of success. P.S.: I would take the report with a pinch of salt as it seems be a bit self-serving 😛, “Even with AI, you’ll still need BCG to solve your business problems” (which is true IMO)!

6

1 Comment

The tempering of GenAI hype Why are analysts so slow and cautious to call it out? David Linthicum brings up a point that has been evident for well over a year - CIOs are nowhere near as enthusiastic as Satya Nadella or Sam Altman about the value of LLMs. Or "Artificial General Intelligence" for that matter. The battleground, as my esteemed colleague Julien SIMON never tires to point out, is Small Specialised Models, and not the bloated, inefficient and planet-eating LLMs. SSMs are being deployed in production, generating value quietly and efficiently. They solve specific problems - semantic search, assisting document authoring, triaging, prioritising, expanding self-help systems. Lots more. 90% of LLM projects are withering in never-ending POC cycles, waiting for magic dust. The other 10% are deployments far away from areas that require precision and compliance. We're really back to what we have long known, which is that Machine Learning is best applied to specific problems that can be solved with proprietary data. "It's the data stupid" writ large. #aiisnotintelligent https://lnkd.in/dFVMunQg

60

5 Comments

McKinsey’s Executive Guide to AI: (Timeline: Why AI now?) The whole AI journey started all the way back in the year 1805 when French mathematician Adrien-Marie Legendre published the least square method for regression which became basis for many machine-learning models. Today  Generative AI (GenAI) is no longer a hype and has gone from science fiction to everyday reality, transforming how we live and work. This revolution kicked off in 2017 when "Ashish Vaswani" introduced transformers, which made huge strides in language processing. Check out McKinsey’s interactive page that highlights the key moments in AI’s journey and shows why it’s such a game-changer: https://lnkd.in/dc6xfqR5 Below is a highlight from the original page. For more details you need to check out the above link by McKinsey & Company. 1805 - French mathematician Adrien-Marie Legendre publishes the least square method for regression. 1958 - Frank Rosenblatt creates the perceptron algorithm, an early type of artificial neural network (ANN), which stands as the first algorithmic model that could learn on its own. 1965 – Birth of deeplearning by Ukrainian mathematician Alexey Grigorevich Ivakhnenko …………………………………………………………………….. …………………………………………………………………….. 1991 – Opening of WWW 1997 – Increase in compute power drives IBM’s Deep Blue victory over Garry Kasparov 1998 - Brin and Page publish PageRank algorithm 1999 - Nvidia releases the GeForce 256 graphics card, marketed as the world’s first true graphics processing unit (GPU). Early 2000s – Broadband adoption begins among home Internet users 2002 — Amazon brings cloud storage and computing to the masses 2004 – Web 2.0 hits its stride, launching the era of user-generated data 2005 – Number of Internet users worldwide passes one-billion mark 2006 – Hinton reenergizes the use of deep-learning models 2007 – Introduction of the iPhone 2009 –  Andrew Ng and his team at Stanford University uses GPUs to train deep-learning models more effectively 2010 – Microsoft and Google introduce their clouds ………………………………………………………………………………………….. ………………………………………………………………………………………….. 2017 – Electronic-device users generate 2.5 quintillion bytes of data per day 2017 – Google introduces upgraded TPU that speeds machine-learning processes 2017 – Ashish Vaswani introduces transformers improving language processing 2020 – In May 2020 OpenAI releases GPT-3 #AI #GenerativeAI #Innovation #Technology #McKinsey #AITimeline #BusinessTransformation #TechLeadership

13