Sharon Zhou, PhD

Quick tip - DDL for vector indexes https://lnkd.in/gwtNtE_r

38

Unapologetically Technical's newest episode is now live! In this episode of Unapologetically Technical, I interview Hubert Dulay the author of Streaming Data Mesh and Developer Advocate at StarTree. We talked about his early experience with web backends like CORBA and SOAP and how those prepared him for data work. He shares his advice for those with web development skills to transition into data and what it's like for a person leaving a company after a long tenure there. We discuss his time at Cloudera and Confluent and how the data industry has changed over time. We talk about his recent experience with Flink and StarTree. We go in-depth into streaming data mesh, covering what it is and how to create real-time data products. We discuss the ways teams can create joined or enriched data products across domains. We round out the discussion with some tips on how to get data mesh adoption. Watch the full episode here:

10

1 Comment

Excellent article by Isabelle Bousquette about spreadsheets and how their versions can quickly proliferate. Any leader focused on improving productivity has encountered this challenge - many versions of a spreadsheet and each has a slight variation. It also highlights the sharp contrast between what enterprises can achieve with AI and what the ground reality is with regards to data - thus what they are able to? And most importantly it brings forth the challenges with tabular data in enterprises. This problem does not occur with images - when has someone opened an image and changed a pixel value or a few of them and stored a version of it ? Tabular data is messy, complex and is the bread and butter of enterprise.

14

4 Comments

AI isn't just predicting anymore—it's creating. Generative AI has exploded onto the scene, crafting everything from code to images with uncanny human-like ability. As transformer models and multimodal systems push boundaries, we're seeing unprecedented automation potential across industries. But with great power comes great responsibility. The future of work and creativity is being rewritten. How will you adapt and innovate in this new AI-powered landscape? #GenerativeAI #FutureOfWork #AIEthics #TechInnovation #DigitalTransformation What's your take on the generative AI revolution? Share your thoughts below.

7

2 Comments

In this article we discuss in simple terms the various strategies for 'breaking up the input' which significantly impact (and complicate) data prep for LLMs. Retrieval-Augmented Generation (RAG) is essential for equipping your Large Language Models (LLMs) with the most relevant enterprise data. But it is way more nuanced than the just 'talk to your PDF' youtube videos which make it look trivial.

4

This new little product demo is awesome! Our new Rev AI Transcript Assistant will save you TONS of time. Easily pull out nuggets of info from your transcripts, making valuable content more accessible than ever. https://lnkd.in/gzzFpcDa #transcription #asr #accessibility #transcript #summarization #ai #assistant

31

2 Comments

AI memory is fundamental for conversational AI and AI agents. It is critical for building personalized experiences and improving reasoning. Check out Letta’s course to understand how to use AI memory to enhance the customization and quality of your AI app.

15

Massive warning about AI incoming, please note that LLMs can’t perform the following tasks - count things, - do math, - logically reason. They are designed to generate human-like text based on the input they receive, but they lack the true understanding and abilities that you possess. This means: - They WILL be wrong if you ask them to do those things. - They MAY ALSO provide incorrect or misleading information. - They CAN produce biased or harmful content. - They CANNOT replace human expertise in critical areas like medical diagnosis, legal advice, or financial planning. - They SHOULD NOT be used as the sole source of truth for decision making. Please approach AI with a clear understanding of its limitations. This starts by first understanding them. Then to build trust in what you’re building/using, do ensure proper oversight and ethical considerations when integrating these technologies into your processes. #AI #TechWarning #LLM

17

5 Comments

A good read on building with LLMs https://applied-llms.org/

11

AI isn't just about replacing humans—it's about empowering them. 💪 AssuranceIQ, an innovative insurance startup, is pioneering a groundbreaking approach to leveraging generative AI: 1. They're using Large Language Models (LLMs) to score customer conversations. 2. This AI-powered system structures unstructured data from interactions. 3. The goal? Enhance the performance of human sales and support teams. 4. It's not about AI vs. humans, but AI amplifying human capabilities. 5. This could revolutionize how we train and develop customer-facing teams. This use case showcases the true potential of AI in business: not as a replacement, but as a powerful tool to augment human skills and drive performance. What other innovative ways could we use AI to empower rather than replace human workers? #AIInnovation #CustomerExperience #FutureOfWork #BusinessAI #TechTrends Share your thoughts on AI augmentation in the comments below!

10

1 Comment

Little-Known AI Insight: One of AI’s untapped potentials lies in multimodal learning, where AI combines insights from diverse data types—like visual, textual, and auditory data—to make far more nuanced predictions and decisions than single-mode AI systems. For example, imagine an AI that doesn't just analyze medical images but also combines patient records, lifestyle factors, and even genetic data. This level of integration could lead to hyper-personalized medicine that we can only begin to imagine today. Although we're seeing progress, few in the industry realize just how transformative multimodal AI could become once these systems fully develop. It’s a frontier that could revolutionize everything from healthcare to predictive analytics in business. Curious to see where this might lead? #ArtificialIntelligence #FutureTech #AI

2

We are just at the beginning of the AI led digital transformation! Lots more questions than we answers on adoption, evolution in Enterprise AI but that’s exactly what creates white spaces, opportunities for innovation. Let’s play!

8

Menlo Ventures has a new $100M fund in partnership with Anthropic called Anthology. Why is it interesting? The 5️⃣ key areas they are focused on investing in: ▶ AI infrastructure—including DevX tools and middleware—that advances the use of Claude and other LLMs, accelerating their adoption and capabilities. ▶ Frontier/novel applications of AI that will transform bio, healthcare, legal, financial services, supply chains, cloud infrastructure, and cyber security. ▶ Consumer AI solutions that transform the user experience by seamlessly integrating LLMs as core components driving interactions. ▶ Trust and safety tooling (at or above the model layer) that enhances AI safety and offers scalable, profitable solutions for responsible AI deployment. ▶ AI apps and tech that maximize societal benefits, such as new forms of education/training, job creation solutions, and tools that expand accessibility. If you had to guess... which of these 5 do you think will make up the largest share of the fund? Extra credit if you add why 🙂

9

This is very topical given the amount of hype and noise around LLMs.

4

Originally at Fortune -- a good (rare) article on neurosymbolic AI even if failing to mention our work at KYield, Inc. We employ a variant of neurosymbolic AI in the KOS (EAI OS), and what I'll call a new exotic form in our SGM. (See my EAI newsletter on 'The power of neurosymbolic AI' - https://lnkd.in/gVGtJTgP). Some good quotes: “Neurosymbolic systems can reach 100% precision, which means being able to fully trust the answers [they give] as accurate, and be fully auditable with every influence on the answer being visible to the customer,” said William Tunstall-Pedoe, a key creator of what became Amazon’s Alexa virtual assistant. “For many applications, such as health care or finance, this can make the difference for adopting the technology at all,” he said.  “No enterprise [AI] app will succeed if you do not have a neurosymbolic engine running, because no customer will be able to trust it,” said Chang. “There are substantial environmental benefits to neurosymbolic AI that come from shifting a large percentage of the processing from very large deep learning models running on GPUs to much more efficient symbolic processing typically on CPUs,” said Tunstall-Pedoe. “This portion of the processing is many orders of magnitude more energy efficient. With models like OpenAI’s o1 doing far more processing than its predecessors to produce results, there is also a continued trend of LLMs increasing their GPU usage rather than becoming more efficient.” Unfortunately, the end of the article gets lost in confusion, making it sound like LLMs are mature science and symbolic is experimental, when precisely the opposite is true. Symbolic has been employed for decades for precision AI. When done well, as we do in the KOS, neurosymbolic is a hybrid of rules-based symbolic logic and accurate data structure, system-wide governance and security, and various algorithms run within a single, highly refined, rules-based system (EAI OS). Multiple methods are used to create and protect high quality data. Journalists -- reach out to me if you want to do a good story on related topics. We've been working in this tech for nearly 3 decades (applied and later theoretical R&D, invention of two complete AI systems, through commercialization) .... and have done deep-dives with dozens of the world's leading organizations. Thousands of scientists are working on various types of neurosymbolic AI -- probably 10s of thousands of data engineers. It's just not hyped like LLMs and GenAI in media... https://lnkd.in/gtMcHkQ4

9

2 Comments

Andrej Karpathy was the director of AI at Tesla and one of the founding members of OpenAI. In a recent podcast, he talked about his new AI+Education company, "Eureka Labs". Here are some (cleaned-up) extracts: "I’ve always been an educator, and I love learning and teaching. It’s something I’ve been passionate about for a long time. One of the broader ideas driving me is the current activity in AI. I feel that a lot of it is focused on replacing or displacing people, which concerns me. I’m more interested in how AI can empower people rather than sideline them." "On a high level, I’m on Team Human—I want to see AI helping people to become even more amazing than they are today. I’m particularly excited about the potential of AI to serve as the perfect tutor for any subject. Imagine how far someone could go if they had access to a personalized, ideal curriculum. We’ve seen that wealthier individuals with personal tutors often excel far beyond their peers. AI could democratize that kind of opportunity." "There’s clear evidence to support this, dating back to the 1980s, with research like Bloom’s two-sigma problem, which shows that students receiving one-on-one tutoring perform significantly better than those in traditional classroom settings. I think AI could replicate, or even surpass, this effect." https://lnkd.in/eU7wzz8G

2

This Thursday November 14 at 11 am EST, do not miss it! I'll be talking live with AI experts Jay Alammar and Joshua Starmer PhD! It'll be a live Q/A, so submit your questions on this thread, and we'll answer them! And we'll also talk about Jay's recent book on LLMs: Hands-On Large Language Models! https://lnkd.in/enXvEdSH

53

5 Comments