Angel Gonzalez

This is a big step forward in the journey from quantum utility to quantum advantage. It’s really interesting to see IBM thinking outside the box, or more accurately off the surface - with a 3D torus approach to connecting qubits. The next couple of years will be the most exciting!

12

𝗪𝗵𝗮𝘁'𝘀 𝗵𝗼𝘁 𝗶𝗻 𝗔𝗜 - 𝘄𝗵𝗮𝘁 𝗰𝗮𝗺𝗲 𝗳𝗿𝗼𝗺 𝗜𝗕𝗠, 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗮𝗻𝗱 𝗚𝗼𝗼𝗴𝗹𝗲 𝗯𝗶𝗴 𝗮𝗻𝗻𝘂𝗮𝗹 𝗰𝗼𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲𝘀 𝗶𝗻 𝗹𝗮𝘀𝘁 𝟮 𝘄𝗲𝗲𝗸𝘀 The past two weeks have been full of innovation in the AI industry, with major announcements from IBM, Microsoft, and Google during their annual conferences. The highlights below reflect my own observations and common themes across vendors. I can see few emerging trends, showcasing the industry's focus on: ✅ Open-source Innovation ✅ AI Agents / Assistants ✅ Small Language Models ✅ Custom AI Chips 1️⃣ 𝗜𝗕𝗠 𝗧𝗵𝗶𝗻𝗸 : 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 𝗙𝗼𝗰𝘂𝘀 𝗮𝗻𝗱 𝗡𝗲𝘅𝘁-𝗚𝗲𝗻 𝗔𝗜 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁𝘀 https://shorturl.at/fk7Yf 🔹 IBM released its Granite family of foundation models into open-source. Developers can now use and modify these models under Apache 2.0 license, 🔹 InstructLab Model Alignment: Partnering with Red Hat, IBM introduced InstructLab, a groundbreaking model alignment technique, integrating community contributions into LLMs, streamlining development process. 🔹 AI Assistant Builder unveiled within Watsonx Orchestrate platform simplifies the creation of custom AI assistants tailored to specific needs, accelerating enterprise AI adoption. 🔹 Strategic Partnerships: IBM announced extended partnerships with industry leaders like AWS, Adobe, Meta, Microsoft, Salesforce, SAP, Palo Alto and others. This includes Allam, an LLM from SDAIA (Saudi Data & AI Authority) which is already available on the Watsonx.ai platform. 2️⃣ 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗕𝘂𝗶𝗹𝗱: 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 e𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗦𝗟𝗠𝘀 https://shorturl.at/ZFXC2 🔹 Copilot AI Agents and GPT-4o: Microsoft's focus revolved around Copilot Studio and their partnership with OpenAI. GPT-4o is at core of multi-modal agents spanning text, speech, and vision. 🔹 Small Phi-3 Models: Microsoft introduced the Phi-3 small models ready to run on end-user devices. It aligns with their Copilot+PC initiative, extending AI assistant experience to personal computers and tablets with Windows. 🔹 Azure Cobalt Chips: Azure Cobalt ARM-based processors are complamenting previously announced Maia custom AI chips, helping to reducing dependence on external hardware vendors. 3️⃣ 𝗚𝗼𝗼𝗴𝗹𝗲 𝗜/𝗢: 𝗦𝘁𝗿𝗲𝗮𝗺𝗹𝗶𝗻𝗲𝗱 𝗚𝗲𝗺𝗶𝗻𝗶 𝗺𝗼𝗱𝗲𝗹𝘀 𝗮𝗻𝗱 𝗧𝗿𝗶𝗹𝗹𝗶𝘂𝗺 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗺𝗲𝗻𝘁 https://shorturl.at/J0ah6 🔹 Gemini Flash, a lighter version of Gemini 1.5 Ultra featuring a 1M token context window (soon up to 2M) offers tp developers better performance and lower costs. 🔹 Gemini Nano - emphasizing the potential of AI runing locally on mobile devices, Google introduced their smallest language model specifically designed for on-device AI applications. 🔹 6th generation of Trillium TPU promises a 5x performance improvement for AI workloads within data centers, and reduces reliance on Nvidia GPUs. We can expect more innovation when clients will be moving from testing and pilots to production.

5

#IBM's #Labradorite and #Merlinite models not only #outperformed #chatbots aligned on human-generated data, but also models aligned on significantly more synthetic data, including #Microsoft's #Orca-2_chatbot, which was trained on 15 million instructions generated by the behemoth #GPT-4 model.

1

VentureBeat writes about Microsoft's Phi-3 model enabling new generative AI applications on-device with its high quality outputs despite the relatively small model size https://lnkd.in/eqR3swC4. #microsoft #phi3 #generativeai #largelanguagemodel #artificialintelligence #venturebeat

6

It won't be surprising if #IBM uses this patent to support enterprises that are building their own #AI_models, offering a #less_resource-intensive method compared to collecting authentic user data.

46

1 Kommentar

It won't be surprising if #IBM uses this patent to support enterprises that are building their own #AI_models, offering a #less_resource-intensive method compared to collecting authentic user data.

21

1 Kommentar

IBM THINK is happening now in Boston…. Enterprise Customers are firmly supporting a number of key priories regarding AI. 1). TRUST is critical. Both the data used to train models and the governance used to ensure they act consistently in accordance with policies, intent and regulatory environments. 2) OPEN provides critical flexability to use the right model - for the right use-cases. 3) ECONOMICS matter - to go from pilot to production requires both a compelling aggregate business case and efficient inferencing unit-costs.

26

On a sunny Tuesday evening on 18 June, a very relaxed #SnowflakeUserGroup Meetup was hosted by Manuel Ifland & team in and outside the synvert Data Insights office in Munich. The group discussed the numerous innovations - especially in the area of #ML and #AI - that were announced and showcased at this years #SnowflakeSummit in San Francisco. How to combine Snowflake's latest #Geospatial features with newly gained insights from ML & AI for #NextGen #BusinessIntelligence or data apps was the scope of my talk with two live demos, including:  ➡️ Acquiring data from the Snowflake #Marketplace and from external storage  ➡️ Cleansing and transforming event and time series data using Snowflake's new set of #H3 SQL functions  ➡️ Training of a forecast model and predicting results with #SnowflakeML  ➡️ Using the instruction following #CortexAI #LLM Complete Function to classify customer sentiment and extract the underlying reasons for that sentiment derived from a customer reviews table  ➡️ Visualizing and exploring Discrete Global Grid #H3 data with #Streamlit in Snowflake (SiS) Unfortunately you couldn't make it to the Meetup in Munich and want to find out more? No problem, give this hands-on lab a try: Geospatial ML and AI using Snowflake Cortex - https://lnkd.in/egXJdMcG I'm already looking forward to the next Snowflake Meetup here in Munich or elsewhere in Germany: with practical examples, exchange of new ideas and best practices from real-world projects and the latest news from our data superheroes. #Snowflake #TechMeetUp #synvert #SnowflakeUserGroup

24

2 Kommentare

How #instructlab #rhelai #openshiftai and #watsonx fit together (along with #podmanailab) to complete an opensource family of AI services. #redhat #ibm https://lnkd.in/g9NqnT-Q

3

Great News! IBM has made its Enterprise-ready Granite Large Language Models (LLMs) to fulfil the promise of "Open" and "Trusted" Generative AI solutions for the Enterprises in the form of IBM watsonx. IBM watsonx Granite Large language Models (LLMs) deliver higher performance at lower cost as compared to other opensource and proprietary LLMs of similar sizes. Find out more in the IBM watsonx post below. NexaQuanta is proud to be an IBM business partner to help our enterprise customers of all sizes in their Generative AI transformation journey Responsibly, Safely, and Cost-effectively powered by IBM watsonx. Zahra Shah David Moore Ian Hepworth Kausar Qazilbash Aarti Samani Chris Donnachie Christopher Jackson Jared Cary Rachel Hicks Barbora Ivanecká Tom Soane Mark Osborn Neil Cornish #ibm #ibmwatsonx #ibmgranite #llm #genai #opensource #aitransformation #nexaquanta

19

1 Kommentar

We recently welcomed CNN’s Richard Quest to IBM Research for a behind-the-scenes look at our IBM Quantum labs and our next-generation IBM Quantum System Two. We discussed the future of computing and IBM’s role in defining technologies like AI and quantum computing. On AI, Richard and I discussed the importance of building models that will actually lead to competitive advantages for enterprises. We also spoke about our big bet for the future — quantum computing — and IBM’s indisputable leadership in driving the world toward the era of quantum utility. You can catch the full segment below ⬇️ https://lnkd.in/d-E6Ba58

886

18 Kommentare

"Organizations are facing an explosion of apps, APIs, events and data spread across hybrid cloud environments worldwide. In fact, IDC estimates that 1 billion new applications are expected by 2028 due to the rapid emergence of generative AI1. As organizations continue their digital transformation journeys, application and data integration solutions are critical for application modernization and effectively deploying AI across the enterprise." These acquisitions of webMethods and StreamSets make sense! Welcome to IBM! #ibmAutomation #AI #watsonx #data #IBM #GenAI #AIstrategy #ibmwatsonx Leon Butler, Labros Kisouras, Camilo Rojas, Larry Baird, Melissa Long Dolson, Nicholas Flood, Shannon Harris, Eunjoo Lee, Ph. D., ROY KOSASIH, Raymon F., Catherine Lian, Diep Pham, Anothai Wettayakorn, Aileen J., Dickson Woo, Viswanath Ramaswamy, Sandip Patel, Mimi Poon, Nelson Lee, Mark Jones, Marius Merkel, Siddhesh Naik, Soojung Lee , Maria Singson, Madhavan Vasudevan, Shanker V Selvadurai, Chetan Krishnamurthy, Priscilla Kim,  Bharati Muthu, Charles Kiang https://lnkd.in/gFAz7Ez5

51

1 Kommentar

#IBM®_watsonx.ai™ is part of the IBM watsonx™ AI and data platform, bringing together new #generative_AI (GenAI) capabilities powered by foundation models and traditional machine learning (ML) into a powerful studio spanning the AI lifecycle. #PowerVS #watsonx.ai ​

4

ZDNet writes about IBM open-sourcing their Granite large language model - doing it the right way also releasing the data, etc. https://lnkd.in/ekXX8RHz. #ibm #granite #largelanguagemodel #opensource #artificialintelligence #zdnet

2

AI-generated code can behave in unexpected ways if developers don't thoroughly review it! Dynatrace CTO Bernd Greifender explains six key reasons why AI-generated code could make it more difficult to maintain the reliability and security of software over the long term

2

🤖 Looking to build chatbots that are not just smart, but secure? Join Patrick Masi-Phelps from Dataiku and Carsten Weidmann from Snowflake as they demonstrate the rapid development of robust chatbots using #GenAI during our webinar in 7️⃣ days. 🙋‍♂️ Don’t miss the live Q&A session where you can ask your burning questions and get real-time answers from our experts. 🔗 Join us to push the boundaries of what GenAI can do for your organization: | https://bit.ly/3wleV19 |

50

#Opensource is vital to the future of #AI #innovation :: In his keynote at Think 2024, IBM SVP and Director of Research Dario Gil asserted that "the future of #AI is Open!!" and outlined the latest technology breakthroughs from IBM that will allow enterprises to scale AI. Watch as he explains in detail what makes for a winning AI business strategy, and why the open-source community is so important to the future of AI innovation. https://lnkd.in/gGpJndde

13

#IBM's Granite #AI models are being #opensourced (Apache 2.0 license) Key points: - The models range from 3B to 34B parameters and are available from #huggingface - Granite models were trained on high-quality data from 116 programming languages, natural language, and code repositories, adhering to IBM's #AIEthics principles. - The released Granite code models are decoder-only models and as such will be primarily useful for generative rather than discriminative tasks. While the Granite code models have a great deal of potential, realizing their full value in an enterprise setting would likely require the base models to be fine-tuned on a company's own codebase and programming practices, and to be integrated into the development tools and platforms that an enterprise already uses, such as IDEs, code repositories, and CI/CD pipelines. It will be interesting to see whether IBM chooses to release more sophisticated encoder-decoder models in the future as the technology matures and enterprise use cases become more demanding. #enterpriseai https://lnkd.in/eZKV8jZ5

3