John Larson

Great to join GAI Insights' call today... and discover this resource on small language models. SLMs will accelerate domain-specific use cases for GenAI. Thank you Luda Kopeikina, PhD in AI, MIT Sloan Fellow and team for hosting me on this morning's call. If you want to understand what GenAI means to your business, I recommend subscribing to GAI Insights content. My take: small language models - commonly defined as LLMs with fewer parameters - offer a simpler and more cost-effective way for companies to apply GenAI to their own data. This matters because GenAI only gives companies a competitive advantage if they apply it to their data. And domain-specific SLMs, such as those highlighted in yellow here, make this easier. They do so because they are trained on the vocabularies and semantics of a specific domain - and nothing more. Picture a model that does nothing but think and talk like a doctor. A hospital chain can take a model like this and use retrieval-augmented generation (RAG) to interpret nurse's notes, diagnose issues and prescribe treatments -- all with careful inspection and oversight by a human expert. What do you think? Below are excerpts from the article. And hats off to the authors, listed below. "Large language models (LLM) have demonstrated emergent abilities in text generation, question answering, and reasoning, facilitating various tasks and domains. "Despite their proficiency in various tasks, LLMs... face limitations due to large parameter sizes and computational demands, often requiring cloud API use which raises privacy concerns, limits real-time applications on edge devices, and increases fine-tuning costs... "LLMs often underperform in specialized domains such as healthcare and law due to insufficient domain-specific knowledge, necessitating specialized models. "Therefore, Small Language Models are increasingly favored for their low inference latency, cost-effectiveness, efficient development, and easy customization and adaptability. "These models are particularly well-suited for resource-limited environments and domain knowledge acquisition, addressing LLMs' challenges and proving ideal for applications that require localized data handling for privacy, minimal inference latency for efficiency, and domain knowledge acquisition through lightweight fine-tuning. "We propose defining SLMs by their capability to perform specialized tasks and suitability for resource-constrained settings, setting boundaries based on the minimal size for emergent abilities and the maximum size sustainable under resource constraints. "For other aspects, we provide a taxonomy of relevant models/methods and develop general frameworks for each category to enhance and utilize SLMs effectively." Fali Wang Zhiwei Zhang Xianren Zhang Zongyu Wu @Tzuhao Mo Qiuhao Lu, Ph.D. Wanjing Wang Rui Li Junjie Xu Xianfeng Tang Qi He Yao Ma Ming Huang, PhD SUHANG CORPORATION WANG Paul Baier Adam Rappaport John Sviokla #ai #llm #genai

33

5 Comments

Business model innovation around Retrieval Augmented Generation from O'Reilly -- they have indexed all the content on Safari and are using it to provide answers grounded in vetted and credible content (their unique sauce). Another innovation here is to credit/compensate authors for the fraction of their work used in a generated answer. While this is progress, the more well-grounded the summaries look, the more dangerous the inevitable mistakes will be -- irrelevant documents will still lead to misleading summaries, and there will still be hallucinations because of interpolations with the text used to train Llama. At least based on the article, I don't see them employing self-checks or guardrails. I hope those are coming. https://lnkd.in/gqqHiR2D

42

1 Comment

LLM Trends - Anthropic is making waves in the LLM market! Anthropic doubled its enterprise presence from 12% to 24% in 2024, as many organizations transitioned from GPT-4 to Claude 3.5 Sonnet, which is now considered state-of-the-art. When choosing a new LLM, enterprises cite security and safety (46%), price (44%), performance (42%), and expanded capabilities (41%) as key drivers. As the LLM market grows 6x year-over-year to $13.8B, Anthropic’s momentum signals a major shift in enterprise priorities. Excitingly, the latest Anthropic Claude models are available through Amazon Bedrock, making it even easier for organizations to adopt. Read the full Menlo Ventures report here: https://lnkd.in/gxnraKfW #AWS #AmazonBedrock #GenAI

21

After the launch of the PCI Data Cooperatives Report (https://lnkd.in/g_CHSvdd) on September 16, the journey continues. Yesterday I spoke on Data Cooperatives and the Digital Commons at the Harvard Club of NYC, as part of the Leaders of Regeneration series. 🌍 As we face climate change and biodiversity loss, and ecosystem collapse, threatening 40% of lives and destabilizing economies, data cooperatives offer a powerful solution. By fostering collaboration and responsible data sharing, they can drive the innovation and collective action needed to protect ecosystems, enhance economies, and create a more sustainable future. #DataCooperatives #DigitalCommons #DigitalGovernance

32

Today's technology disrupters are examples of regulatory entrepreneurship (OpenAI, Uber, Spotify, etc.). The AI revolution, with GenAI, has disrupted the momentum of crafting and implementing regulations.  Now, new entrepreneurship will focus on leading and implementing regulations and independent audits. #ai #airegulation #genai #generativeai #aiml #aiethics ##responsibleai

11

Thanks to #genai360 for the shoutout here for Consilience AI and our AlphaIQ product. Davit Buniatyan Joseph Byrum, Ph.D., MBA

10

This is a nice breakdown on some of the emerging ways AI are being use to create new businesses. I look forward to studying the emerging pricing models that these companies will be adopting, particularly on the “AI colleagues” and the “AI native services” categories.

8

Reading this article made me think of Matt Webb and Michael Veitch, the great guys who run OutcomePath (the pre-spend agency). If you’re considering how to plan and prioritise tech implantations, including AI, at your organisation, give Matt and Mike a shout and they’ll help you deliver better, quicker and with less waste. https://lnkd.in/evjcvAnQ

15

1 Comment

"When AI is paired with human knowledge, experience, and perspective, it moves beyond its role as a tool. Together, the duo unlocks AI’s processing capabilities with the nuanced understanding and intuitive insights that only human professionals can provide. It is here where the true value of AI emerges." #humanaiconnection #aioptimization #polarcomplementarity #aiage

2

"When AI is paired with human knowledge, experience, and perspective, it moves beyond its role as a tool. Together, the duo unlocks AI’s processing capabilities with the nuanced understanding and intuitive insights that only human professionals can provide. It is here where the true value of AI emerges." #humanaiconnection #aioptimization #polarcomplementarity #aiage

5

Adrienne Murphy discusses an outstanding example of data leadership by the Massachusetts Department of Early Education and Care. Transcript: Adrienne Murphy (Massachusetts Department of Early Education and Care): First, in terms of understanding the pre-K landscape, the CUSP network has really helped us understand, in this three and four-year-old space, who's being served across the mixed delivery system, which has been especially important when our governor, the Healey Administration, announced a new Gateway to Pre-K Initiative, which included a proposal to deliver universal, high-quality preschool access for four-year-olds in all of our Gateway cities, which are our sort of mid-sized urban centers that are sort of the regional economic anchors across the state. And the cost model has been able to help us to benchmark sort of where are we now, who's being served across our licensed and funded early education care providers, across our public K-12 system, across our private school system, to be able to understand sort of the full landscape of who we're currently serving. And, by virtue of who we're currently serving, also, who are we not serving? Sort of what's the delta between where we are now in pre-K access for four-year-olds across the state, but really particularly in these gateway cities? And then sort of, what's the delta between where we are now and the universal access that is proposed in the gateway to Pre-K initiative? So, this is, you know, really one instance in which being able to know who are we currently serving, who are we not serving, and where are we not serving children has become really key for us and for Massachusetts as sort of the policy design process. Click here to learn more: https://team3si.com/ #communityofpractice #dataforgood #earlychildhooddata #acf #ecids

1

Introducing Beyond-AI’s Solution for Reliable, Near Bias-Free Insights – Elevating LLM Performance. In a world where both AI-driven searches and traditional search engines often deliver biased, outdated, or misranked results, Beyond-AI’s advanced ranking system offers a better way. By feeding LLMs with precise, credible, and up-to-date information, Beyond-AI Matrix Pinnacle enables LLMs to reach their full potential, free from the noise and inaccuracies that can otherwise hold them back. How it works: Beyond-AI allows users to run unlimited search processes in parallel, with each search leveraging 70 distinct search services simultaneously. For each selected URL, Beyond-AI reads the full first page and applies a sophisticated, proprietary algorithm that ranks relevance with unmatched accuracy. The system also integrates data from relational databases, NoSQL, digital content, vector data, and more, packaging all sources in embedded form for the LLM to process with clarity and precision. This ensures LLMs receive only the most credible, relevant data, vastly enhancing the quality and reliability of their responses. Why it matters: With Beyond-AI as the foundation, LLMs overcome limitations like outdated or incomplete data and deliver more intelligent, well-rounded responses. Beyond-AI empowers decision-making with dependable insights, redefining the search experience by letting LLMs truly thrive. With this level of knowledge and precision, why would you ever run a search the old way or ask an LLM a question without Beyond-AI empowering it? Tom Brne [email protected] #BeyondAI #ReliableAI #LLMEnhancement #MisinformationFree #NextGenAI #AI #AIInnovation #NextGenTech #IntelligentSearch #AdvancedAI #LLMOptimization #DataPrecision #SmartAI #AIEnhancement #ReliableInsights #FuturisticAI

1

Is an individual Enterprise LLM License a justifiable expense for your team members? I've seen this logic circulating: "If someone earning ≥$62,400/yr can save just one hour in a month using a #GenerativeAI tool, it justifies the cost of a $30/month #EnterpriseLLM." This assumes time directly equals productivity, and that a user doesn't already have an hour+ of free time in a month. We know this isn't always the case. Two quick factors to look at first: 1. The litmus test: Will time saved be reinvested into productivity? Is that productivity worth more or less than an hour of time on paper? 2. This simple formula: S>=max(C/H,F) In words: The number of hours saved (S) must be greater than or equal to either C / H (cost divided by hourly rate) or F (existing free time), whichever is larger. For the record, if we're using these tools effectively, an hour saved each month for is on the extreme low end of what I would expect for most business roles...

8

9 Comments

A really practical guide to LLMs is now available! Learn what to do, what to avoid, from prompting to RAG, to fine-tuning, and evaluation. Definitively worth reading. https://lnkd.in/dCC6uKjs

34

The USPTO team integrated Sonatype solutions and revolutionized their approach to software development. By automating the critical aspects of the #softwaredevelopmentlifecycle, they’re not just keeping pace with technology—but leading the charge in IT innovation. Curious about their journey and how we're tackling challenges with #opensource dependencies head-on? Dive into our story 👉 https://bit.ly/4aSrCz2

8

A few days ago, Dario Amodei published an essay expressing his vision for #AI with keen insight and wisdom. Techie or non-techies, PLEASE READ: https://lnkd.in/gRBv_6E6

4

4 Comments

Ever paused to wonder if... Technology understands you? Not superficially, but the real you? That's the spark behind Kwaai Tech innovation that gets you. Made accessible, open, and beneficial for everyone. From poets to programmers, embracing diverse thinking hats.  Lasered focused on individuals' privacy and rights. Kwaai is more than a project; It’s a movement towards ethical, personal, and inclusive AI. Every skill, every voice, can help steer the course. Let’s make AI personal. The Kwaai way. #JoinKwaai #AIForUsAll #EthicalTech

4

1 Comment