Igor Perisic

Good article on diminishing improvements in AI foundation models. Lack of training data is a main culprit — "But compute needs might not actually be the largest barrier when it comes to improving A.I. model capabilities, according to the venture capital firm. It’s the availability of training data needed to teach A.I. models how to behave that is increasingly becoming a problem. https://lnkd.in/ein7s2JE

66

4 Comments

Great initiative by Cohere to bring better tools to expand LLM coverage to non English languages. From my Google Translate work, I've learned about huge disparity in training data for different languages: while only 20% of world's population has some skills in English language, 50% of Internet content is in English. That means that for many languages (and especially Indic languages that are spoken by ~20% of world population but duesnt have much of Internet content), shortage of training data could be significant inhibitor to LLMs quality. #artificialintelligence #largelanguagemodels #GenerativeAI #airevolution

10

Lin Guan and Yifan Zhou have led a very interesting study investigating the effectiveness of Vision Language Models in critiquing robot behaviors in terms of the desirability of their "style." If you are wondering how our work on LLM's inability to verify correctness squares with this study on using them as style critics, see my earlier post https://lnkd.in/gBYAr5qq ! This is the reason why being a style critic is one of the roles LLMs can play in LLM-Modulo framework: https://lnkd.in/gyAyKx4E Check out the details of the work @ https://lnkd.in/gfVZ4EYW ( Lin Guan also has a nice twitter thread on the work here: https://lnkd.in/gnieU-qM )

89

2 Comments

From Azeem Azhar on Eric Schmidt’s recent talk at Stanford: “Agent-based systems, which autonomously execute multi-step tasks and adapt to environmental feedback, are poised to unlock significant economic value next. Building on this foundation, text-to-action capabilities will emerge, translating natural language into specific, working programs. Imagine your own personal programmer at your fingertips, ready to create whatever you envision. A dating app that matches people based on their Netflix viewing history and Spotify playlists? A prototype within minutes. These forces will converge. While he doesn’t say it explicitly, it seems that Eric expects the next two years to be faster and more turbulent than the previous two.” https://lnkd.in/gxWEy5Pf

8

I'm thrilled to see Anthropic's latest announcement of the upgraded Claude 3.5 Sonnet and new Claude 3.5 Haiku models! 🚀 As someone who strategically leverages various AI platforms, including Claude, for specific use cases, these advancements are particularly exciting. The enhanced coding capabilities in 3.5 Sonnet are especially promising 💻✨, as they build on Claude's already impressive performance in this area. What really stands out to me is the groundbreaking computer use capability now in beta. 🤖🖱️ While I currently use AI selectively and thoughtfully for specific tasks, features like this - where AI can interact with computers like humans do - hint at the expanding possibilities ahead. I appreciate Anthropic's transparent approach in acknowledging that this feature is still experimental and may have limitations. This aligns with my philosophy of carefully choosing when and how to implement AI tools as they mature. 🎯 The fact that Claude 3.5 Haiku matches Claude 3 Opus's performance while maintaining cost efficiency is also noteworthy. 📊💪 This kind of advancement helps organizations like ours make informed decisions about scaling our AI usage as the technology evolves. Looking forward to exploring these new capabilities! 🌈 #AnthropicAI #Claude35 #AIInnovation #ComputerUse #FutureOfTech

2

Introducing Claude 3.5 Sonnet: Revolutionizing AI with Speed and Precision Claude 3.5 Sonnet represents a significant leap in AI, offering advancements in speed, precision, and capabilities such as humor understanding and handwriting recognition. Its introduction could accelerate AI development and application across various fields. #claude #ai #sonnet

🚀 Excited to join the SaaStr AI Summit in San Francisco this week! 🚀 💡 If you are interested in talking about how we can leverage LLMs to generate empathetic and intuitive conversations to create 𝐚𝐜𝐭𝐢𝐨𝐧-𝐨𝐫𝐢𝐞𝐧𝐭𝐞𝐝 𝐮𝐬𝐞𝐫 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞𝐬, DM me. 🗃️ There are many companies in the AI space focusing on developing AI Agents using RAGs to replace FAQs or provide more intelligent conversations. 💬 However, our vision is to develop 𝐂𝐨𝐧𝐯𝐞𝐫𝐬𝐚𝐭𝐢𝐨𝐧-𝐀𝐬-𝐀-𝐒𝐞𝐫𝐯𝐢𝐜𝐞 𝐛𝐲 𝐩𝐫𝐨𝐯𝐢𝐝𝐢𝐧𝐠 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭𝐬 𝐭𝐡𝐚𝐭 𝐜𝐚𝐧 𝐭𝐚𝐤𝐞 𝐚𝐜𝐭𝐢𝐨𝐧𝐬. These systems go beyond simply providing information—they can renew insurance policies, change your payment preferences for a subscription, mail you a letter as a proof of address, handle bank transactions and perform other sophisticates tasks autonomously. 🎬 Of course there is a lot more to these Action Agents than just the functionality. These empathetic conversational agents not only understand emotional cues but can also adjust their tone to fit the situation. This emotional intelligence is crucial in customer service, where addressing frustration with care can lead to increased customer satisfaction and loyalty. ✅ If you want to explore how you can 𝐛𝐮𝐢𝐥𝐝 𝐚𝐧 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭 𝐰𝐡𝐨 𝐜𝐚𝐧 𝐠𝐞𝐭 𝐭𝐡𝐢𝐧𝐠𝐬 𝐝𝐨𝐧𝐞, combining actionable and empathetic capabilities, let's connect. #AI #LLMs #ConversationalAI #ConversationAsAService #SanFrancisco

30

Together AI Introduces Mixture of Agents (MoA): An AI Framework that Leverages the Collective Strengths of Multiple LLMs to Improve State-of-the-Art Quality Quick read: https://lnkd.in/gejHtYDp Paper: https://lnkd.in/g8sbqAWn GitHub: https://lnkd.in/gYsmsZa6 Together AI

38

2 Comments

"Productization is an affirmation," writes Adrian Bridgwater, "when any good or service evolves to a point where it can be productized and boxed, it denotes a certain solidity and branded consumability." In our conversation, I was delighted to share how #SAS equips organizations with readily deployable AI technology to productionize real-world use cases. https://lnkd.in/enUGCgAi #AIML #sasusers

9

From where I sit in the SF Bay Area, cradle of innovation for at least the last 30 years, I'm a true believer in the promise of the #AI revolution. That said, I find it fascinating to consider views on the opposing side. I'm sharing below two opinion pieces from the last two weeks suggesting that the AI revolution is fundamentally overhyped and a bubble ripe for bursting, and that AI tools are actually not useful or transformative. Contrast this with venture capitalist Kai-Fu Lee's prognostication that 50% of white collar jobs will be displaced by #AI within three years, and that OpenAI and the other cutting edge movers in the space are headed to a trillion+ dollar valuation. Which side will win out? Where will we be by 2030 in the AI race? I'd love to hear your thoughts in the comments.

7

3 Comments

A must-read! Thanks, Tim Sanders for the recommendation and the book itself! Not just due to their Canada-connection, Ajay Agrawal, Joshua Gans, and Avi Goldfarb packed this ~200-page book with not just facts and figures, but frameworks and fascinating, fantastic food for thought. One key takeaway for me is to cut the use of 'insights' from my vocabulary when it comes to the impact that #AI brings to businesses (insights on the tech and trends are still great though, heck our team is called the Research Insights team). When we think about the power that AI is bringing to enterprises, it is all about its impact on 𝕡𝕣𝕖𝕕𝕚𝕔𝕥𝕚𝕠𝕟𝕤 and 𝕕𝕖𝕔𝕚𝕤𝕚𝕠𝕟𝕤. Saying that it helps businesses generate or gain insights is too surface level and not getting to the core of AI's capabilities. Business leaders, take note: focus on what predictions your businesses rely on and then figure out if, how, and when AI can come and improve them.

30

4 Comments

Who will get crushed by Claude LLM coming to iPhones? Will existing apps (and companies) get steamrolled by this? For those that don’t know Claude is a suite of 3 LLM models: Haiku, Sonnet and Opus. - Opus knocked ChatGPT off the leaderboard on capability when it was introduced. - It can hold a massive 160,000 words “enough for a user to paste in a weighty novel and ask follow-up questions” - Claude is multimodal so you will be able to take pictures or upload images from your iPhone and include that in your prompt Use cases * After a meeting, a business user could snap a photo of a whiteboard diagram and ask Claude to summarise the key points, making it easier to share and act upon important information. * Similarly, a consumer could take a picture of a plant they encounter on a hike and ask Claude to identify the species and provide more information about its characteristics and habitat. ** think the Picture This app ** What other uses do you see for this? https://lnkd.in/gdExfpx9

6

Heading to Southern California! Join me and Garrett Springer at a free public meetup on July 25th where I will talk all about LLMs. (Don't tell corporate, but it's not really about Snowflake) 😃 What I want to talk about: 3 Steps for training a LLM - these have consequences around hallucination, agentic workflows, and expectations of LLMs Improving LLM Performance - I get a lot of questions about training your own model, fine-tuning, prompting, long context, function calling, so lets address them Using LLMs - Importance of evaluation and the latest lessons, Risks around LLMs (security, legal, ethical), and having a strategy to using LLMs If you are in the area, swing by and say hi. I will also make a recorded version of this talk at some point before the end of summer. Meetup Info: https://lnkd.in/g9qmgvtx Check out my talk from March: https://lnkd.in/g6nhnKF7

30

2 Comments

Congrats VERSES (OTCQB:VRSSF) Gabriel René and Dan Mapes and Dr. Karl Friston for introducing an "efficient alternative to deep learning, reinforcement learning and generative AI called Renormalizing Generative Models (RGMs) that address foundational problems in artificial intelligence (AI), namely versatility, efficiency, explainability and accuracy, using a physics based approach." #ai #activeinference #RGM #spatialweb #AGI #cognitivecomputing https://lnkd.in/gcYSM4eN

24

Did you hear? Anthropic's upgraded Claude 3.5 Sonnet model is now on Vertex AI! When building with the upgraded Claude 3.5 Sonnet on Vertex AI, you get a platform to take your workloads from concept to production with fully managed infrastructure, robust security and privacy, and advanced developer tools. Get all the details, here →

14

New on Runtime: The August edition of the Runtime Roundtable is live! We asked our panel of experts to help companies that are setting up platform engineering teams understand how they can maximize that investment and finally (theoretically) realize the promise of DevOps. With thoughts from Yuri Golikov, Sudha Raghavan, Sakib Dadi, Steven Jones, Madhukar Kumar, Patrick Bohannon, Ryan Fox-Tyler, and Hubert Behaghel.

31

1 Comment

"what was missing from language models was “they didn’t understand the spatial context you’re in . . . so that limits their usefulness in the end"...Demis Hassabis Yann LeCun said LLMs had “very limited understanding of logic . . . do not understand the physical world, do not have persistent memory, cannot reason in any reasonable definition of the term and cannot plan." The solution? VERSES The Spatial Web Standards will enable new forms of AI to evolve beyond large language models. Spatial Web Foundation George Percivall Gabriel René What say you Gary Marcus The Standard for Spatial Web Protocol, Architecture and Governance is moving to the final balloting process at the Institute of Electrical and Electronics Engineers (IEEE). IEEE “This progress is the result of our efforts over the past five years where we have been working to develop the social and technical standards with the IEEE to help ensure that AI systems are safe, secure, and trustworthy. The report we issued, ‘The Future of AI Governance,’ combined with the legal input from the law firm Dentons, provides our unique perspective on global artificial intelligence governance. Dentons Link Legal #ai #activeinference

19

1 Comment