Nithin MATHEWS

🧠 The myth: More data makes an #AI model smarter and therefore more accurate 💪 The truth: It’s not just about data—it’s about #parameters With the around OpenAI’s #GPT4o it might seem that feeding more data into these models makes them smarter, the real game-changer is the number of parameters they have like Meta #Llama 3.1 with 450b parameters 💡 More parameters mean these models can process and understand data in more sophisticated ways, leading to increased accuracy and the appearance of being “smarter.” More parameters = higher trustworthiness, even with more focused data like the proprietary #data in your organization In the world of conversational user interfaces becoming the norm, more parameters will be crucial for a seamless experience. While yet too slow and too expensive for real fast interactions between humans and machines, with #quantumcomputing we will move in a new area of responsible #UI #design #AI #MachineLearning #LLM #conversationaldesign #hmi #Innovation

5

1 Kommentar

Apple targets Google staff to build artificial intelligence team - SWI swissinfo.ch https://buff.ly/4biSyYK - 🔍 Apple has been hiring AI experts from Google to expand its global AI and Machine Learning team - 🏢 Apple has a secretive European laboratory in Zurich, Switzerland - 📉 Apple's shares have slipped, putting pressure on the company to announce game-changing AI features - 💬 Apple is focusing on deploying generative AI on mobile devices - 🧠 Apple has been developing AI products, such as Siri, based on neural networks - 🤝 Apple's top AI team includes former key figures from Google - 🧑‍💻 Apple has acquired numerous AI start-ups in the past decade - 📱 The next big thing for Apple could be AI smartphones with more powerful chips - 🤐 Apple has been tight-lipped about its AI plans - 💡 Apple's foray into generative AI features may be revealed at its Worldwide Developers Conference in June

2

1 Kommentar

A few ideas on the role of #GenAI and #LLMs beyond the action of writing code (below). In short: there seems to be a lot more room for improvement in code-adjacent tasks. https://lnkd.in/gbYUtNT3

8

Most discussions about AI remain centered on Silicon Valley, presenting a narrow, often US-focused perspective. But on the latest episode of #BBC #AIDecoded breaks that mold, taking the episode to #Berlin to explore Germany’s vibrant and unique #AIEcosystem. From #healthcare #innovations to #artisticcreativity, the episode highlights how Germany’s approach to AI reflects its cultural values, prioritizing trust, precision, and long-term impact. As the amazing co-presenter Stephanie Hare (and our guide for Berlin) aptly pointed out during the episode: “You can either be the first, or you can do it well.” Germany’s AI ecosystem embodies the latter—focusing on trust and excellence rather than speed. This perspective reminds us that innovation isn’t just about being first—it’s about delivering value that lasts. It’s going from a subject perspective to objective This shift in focus is a crucial reminder that AI’s global story cannot be confined to just the US or UK. Broadening the narrative to include diverse ecosystems like Germany’s reveals new ways of thinking about innovation—ones that challenge dominant paradigms and enrich our understanding of AI’s role in society. How can we ensure the global AI conversation is inclusive and representative of different cultural approaches? #ArtificialIntelligence #InnovationBeyondBorders #TrustInTech #AIDecoded #inclusiveAI #BerlinAI Seán Boyle Ramit Debnath Shannon Vallor Lynne Craig Dr Shama Rahman Yann Mathias Tyler Shores Mara Pometti Wakanyi Hoffman Payal Arora

12

5 Kommentare

The German AI Federal Association welcomes EU plans to establish "AI factories" where young companies can train and test their "artificial intelligences" and language models on supercomputers. This is in response to a vote in the European Parliament on the "EuroHPC Startup Initiative to Strengthen Europe's Leadership in Trustworthy Artificial Intelligence". The Commission's President Ursula von der Leyen believes that Europeans can develop a leading role with a focus on "trustworthy" AI and wants to facilitate access to suitable supercomputers for AI startups. As the EU continues to invest in its AI infrastructure and support startups, enterprises should closely monitor developments and explore ways to leverage (talent and supercomputing resources) the emerging opportunities in the European AI landscape. #ai #enterpriseai #euAI #trustworthyai https://lnkd.in/eNDxMBPp

5

During searching for new ideas for first steps towards superintelligence I came to the intersection between cognitive evolution and LLMs, for example here the evolutionary algorithms (but given enough compute power one could conceive of simulating cognitive evolution and then increasing the simulation speed). I have not yet completely read the article, but it more and more goes in the direction, at least from the way of thinking, of OpenCog's Hyperon (https://lnkd.in/ek3HaNeX), which is of course way more sophisticated. #artificialIntelligence #cognitiveEvolution #superintelligence #openCog

2

Today, the renowned Sepp Hochreiter – Austria’s leading AI expert in academia – mentioned during his keynote at Expedition KI 3.0 that LLMs (Large Language Models) aren’t interesting, and we should focus on “real” AI. I strongly disagree, and here’s why: By integrating LLMs into Multi-Agent Systems, companies can: • Boost efficiency: Automating tasks and increasing productivity. • Enhance decision-making: Analyzing data for strategic insights. • Stay ahead of the competition: Identifying and responding to market trends more quickly. • Cut costs: Optimizing processes to achieve substantial savings. • Personalize customer experiences: Delivering tailored interactions that improve satisfaction and retention. These are just a few of the many benefits. Of course, I suspect he might have been speaking sarcastically! 😉 Visit us at our stand! #expeditionki #expeditionki3.0 #flughafenwien #sunrai

7

4 Kommentare

Efficient. Fast. Autonomous. And one day it will erase humans: #AI I personally always said there is another perspective to artificial intelligence and the only thing that is super is the outcome for humans. Philosopher Nick Bostom has a new book, and it’s finally acknowledging the potential of a harmonious human-AI relationship and its problem solving capabilities. AI = augmented intelligence #design #ai #problemsolving #innovation #creativeai

11

1 Kommentar

💡 The 2024 🇨🇭 Open Source Study confirms that 97% of Swiss corporations are using #opensource software …. and 40% are already using #opensource #ai tools and models. Here my Top 5 takeaways from the study ⬇ 1. There are already 700’000+ open source #ai models on Hugging Face 2. #TensorFlow and #PyTorch are the top used #ai open source tools. 3. #Llama and #Mistral are by far the most used open source #ai models. 4. #opensource is used where nobody can really see it! None of the #cloud environments we’re using in private and professional live would work without Open Source! 5. #linux, #docker and #kubernetes are key technologies to run today’s Cloud environments. A lot of interesting articles about the economics of #opensource and real world examples are complementing the study. Thanks to CH Open, swissICT and Berner Fachhochschule BFH for running an providing this interesting study! Here you can find the details (in German): https://oss-studie.ch/ Matthias Stürmer Alain Gut Christian Keller Richard Zobrist Katharina Göppinger Lise Crettaz Philipp Spaeti

33

4 Kommentare

👉 Greater Zurich: The Global AI Powerhouse Where tech giants like Google, Meta, Microsoft, and NVIDIA lead AI research and innovation 💥 Are you interested in collaborating with AlpsGenAI community members or expanding your network in the Alpine region? Are you considering relocating your company or AI product development to the Greater Zurich Area? Contact us! We help innovative companies scale, innovate, and grow in the Alpine regions. Why the Greater Zurich Area is the ideal place for AI companies to innovate, scale, and succeed: The Greater Zurich Area is a rising global AI hub, powered by Switzerland's expertise in precision mechanics and advanced industries. With world-class universities and top innovation rankings, the region attracts tech giants to drive AI breakthroughs. Key Advantages: - Leading Innovation & Research - Thriving AI Ecosystem - Perfect Launchpad for Growth in the Alpine countries Explore the whitepaper to learn why Greater Zurich is the perfect environment for AI companies to thrive. Switzerland ranks #1 globally as the Best Country in 2024. #ArtificialIntelligence #AIInnovation #GreaterZurichArea #TechHub #AIResearch #AICommunity #InnovationEcosystem #AlpsGenAI #AILeadership #SwitzerlandTech #AIStartups #TechGiants #AIProductDevelopment #InnovationNation #SwissExcellence #AIHub #BusinessGrowth #AITrends #GlobalInnovation #TechInAlps

6

Celebrating New Beginnings 🚀 Exciting times ahead! After years of collaboration, my best friend and business partner, Jonas Aaron Vander, and I are launching a new AI (ad)venture—SunRise AI Solutions GmbH. 🌅 As an AI Innovations Agency, we’re passionate about democratizing AI for Startups. We believe that every business, from small startups to established SMEs, should have a seat at the table in this booming AI economy. 💼🤖 Our mission: 1) to ensure that every company—no matter the size—gets their fair share of the booming AI economy. 2) Offer deep tech solutions beyond the typical AI wrappers. 3) Leverage our extensive network for grants, funding, and AI talent connections. We’re already working with 10+ amazing clients and can't wait to share some success stories soon. 🎉 Let’s chat if you’re looking to supercharge your AI strategy! 👉 What are your thoughts on how AI can help smaller companies compete with the giants?

35

15 Kommentare

Pondering the learnings from the past couple of days in Frankfurt… Key takeaways: 1. Going from pilot to production with #ai is still a challenge. David Schreffler shared how one company he’s spoken to ran 75 #genai pilots this year… 3 went to production. 2. #GenerativeAI has great capabilities, but it’s not the solution to all of your problems. Companies like Citi have reached 350 million annual interactions without end to end gen ai applications. 3. There’s still a big focus on AI being a ‘cost saver’ with ROI coming in the shape of labour savings. In my opinion, this is a fallacy and misses the real opportunity: scale and transformation. 4. Agentic applications are showing promise in complex areas that require reasoning (great talk from Cobus). Making it work in enterprise isn’t straightforward yet. 5. AUTODOC is doing 7m interactions per year in its chatbot with a 64% success rate. Digital interactions across channels up from 24% to 33% because of the #chatbot. Great to see folks like Christoph Esslinger, Ph.D. and the team at Kore.ai : congratulations on a great event. Good vibes, good content, good show 👍 #enterpriseai

43

9 Kommentare

What Karl says. There's going to be two amazing talks about Hopsworks based #ai and #ml systems: * Build a personalized Bitcoin (BTC) virtual assistant in Python with Hopsworks and LLM function calling by Javier de la Rúa Martínez (https://lnkd.in/eQfWpFRU ) * Build TikTok's Personalized Real-Time Recommendation System in Python with Hopsworks by Jim Dowling (https://lnkd.in/eXEHePjG) If you are in the #berlin area next week - go and take a look!

7

A relatively short but inspiring TED talk by Stephen Wolfram late last year surmising, based on my understanding, that the universe is essentially computational at its core but also computationally irreducible. In essence, there are many possible computable pathways through each iteration of time. https://lnkd.in/gFfidMx9 #computation #computational_irreducibility #AI

5

1 Kommentar

As we enter into the 'selling work' software paradigm with AI-enhanced services, it's worth reflecting on the breakneck speed of developments in hardware and model capabilities and what they imply for this new era. - With the new generation of B200 NVIDIA chips, we need 1/4th the power to train a 1.8 trillion parameter GPT-MoE model. - Inferencing performance of Blackwell chips could improve from 5-21x for a ~2x increase in compute costs. h/t Clark Tang - We're going to go from between 50 to 100 tokens p/s to 500 to 2,000 this year. h/t Swyx - Mosaic's Law asserts a model of a certain capability will require 1/4 the [money] every year from [technological] advances. This means something that is $100m today goes to $25m next year goes to $6m in 2 yrs goes to $1.5m in 3 yrs. h/t Naveen Rao - We've gone from 32k context windows to a 1 million context window in one year Google - Context utilisation has increased 4x in one year h/t Louis Knight-Webb With all these advancements, what possibilities are unlocked? We'll be able to innovate on UI/IX, displacing the streaming of tokens we see today. The more profound consequence though is the throughput of work that will be unlocked by asynchronous full-stack employees (h/t Alessio Fanelli). https://lnkd.in/epM7aRwJ

34

2 Kommentare

AI is more than GenAI and Daniel Whitenack explains why. A great new episode of the #PracticalAI #podcast. Not to be missed to understand the AI ecosystem holistically and how models, embeddings, data, prompts, etc. all fit together. In this episode, Daniel - CEO of Prediction Guard - breaks down a history of developments in data science, machine learning, AI, and GenAI to give listeners a better mental model. https://lnkd.in/dgn7gQup #AI #genAI #generativeAI #ArtificialIntelligence #ML #MachineLearning #DataScience #startups #AIstartups #IntelLiftoff #PredictionGuard

20

1 Kommentar

Securing generative AI presents unique challenges beyond traditional security approaches. As generative models, like large language models (LLMs), process diverse multimodal inputs (text, images, audio), they demand tailored safeguards against hallucinations, data vulnerability, and high processing demands. Key considerations include securing training and inference phases, ensuring data privacy across cloud or self-hosted systems, and addressing the complexity of probabilistic outputs. Given the rapid adoption of generative AI, evolving strategies, including red-teaming and compliance, will be vital for safe deployment across industries, as Matt Honea explains in his multi-part series. #ZeroTrustAdvocate

1

Seem familiar? The future is ….. The thing we have really got to do is ….. in this case - The Future is AI. AI is going to be the next biggest way to solve our challenges. Have we not observed similar patterns in the past? Thinking Pharma-based research here: RNA Interference, Gene therapy, Pharmacogenomics, nanotechnology drug delivery systems as examples. They have each proven to have potential benefits and indeed successes are present, even if they have been slower and more difficult to apply generally. This well researched report from Gartner Group, reinforces the tenant that the success of AI is to aim for realistic goals. Understand what data you need to use in the models and most importantly - build data models that are built on highly curated data that is the correct data for the outcome in mind. You might even want to read the last sentence again….. How many failures does it take for us to make us turn our attentions elsewhere? For a scientist working on the basics - depends on the tenacity of the scientist, but probably way more times than we would admit, as we generally do not like to lose. For company executives, on the other hand, where the impact of failure is very high, then we may only get a single shot at this. It is as the old saying goes - Garbage in Garbage Out. AI is a powerful tool, anyone who has used ChatGPT(tm) will know - but when you look deeply at the output it gives, you need to be careful of accuracy and also applicability. As a scientist, also passionate about the potential of AI in the world of pharmaceutical research, analytical data interpretation and so on, the voice in the back on my head keeps saying to me - “learn from the past:” acknowledge that we are very much infants in the world of AI. As a parent, my analogy is the progression from baby to adult. So like infants, we need time to play, time to learn, time to adapt, time to enjoy our successes, time to optimize and deliver. As we grow and we understand, then we can with confidence take the maximum benefits offered be this powerful technology. Explore small and grow big by building on our wins, rather than going all in. So, while we may well see some cooling of interest over the next few years, let’s hope that we can use past lessons and maybe a check on reality from time-to-time to keep us on a road to success. Any views, opinions, and conclusions shared here are my own. I welcome constructive feedback and contributions.

8

2 Kommentare