Anita Nesbitt

See how to use HeatWave GenAI's in-database vector store, LLM integration, and natural language capabilities in an e-commerce application: ⊙ Automatically generate product review summaries. ⊙ Use natural language translations for product descriptions and product review summaries. https://lnkd.in/g2dkqkVk #heatwave #genai #vectordatabase #ai #oraclecloud #oci #heatwavegenai #mysql

24

Some takeaways for a Jeff Hinton interview with timeless insights. My POV as an app builder is that models are now being improved for better UX (faster response/multi-modal) and reasoning capability and cost continues to improve. The technology headroom remains strong, and the integration of LLM into all software experience will continue and expansion of the reach (invasion) of software + intelligence will also continue. - Next word prediction is not merely parroting of training data. To do next word prediction well, LLM needs and already exhibits strong reasoning behavior. This would continue to scale / improve with more data / training. - LLM is a compression of human knowledge. The compression mechanism is identifying similar concepts across domains and reduces duplications among them. This can also be a source of "creativity" as LLM "sees" correlations of seemingly unrelated ideas through this compression. LLM can be more creative as it sees similiarity / correlations across all domains beyond (human) convetional wisdom. - LLM can exceed the quality of its training data. Jeff gave an example of an experiment where a model trained on 50% intentionally incorrectly labeled data, can still perform at 90+% accuracy after training. - Multi-modal models can improve the overall world understanding and lead to better overall reasoning. It does not even need to generate multi-modal output. - Not ML specific, but an interesting life insights. Bad intuition comes from believing (or wanting to believe) everything you are told. It diminishes the ability to make a clear choice that would lead to differentiated success (or failure).

4

So much fun talking about the power of LLMs with Amazon Bedrock and Anthropic Claude. In summary, the ever-evolving landscape of GenAI, it's becoming clear that there won't be a single LLM (Large Language Model) to rule them all. The rise of application-specific LLMs heralds a new era, where various models will carve out niches to excel in. This diversity opens up a wide playing field for innovation and tailored solutions. When it comes to deploying AI models, striking the right balance in evaluation is paramount. We must weigh factors like cost, speed, and precision to ensure optimal performance and resource utilization. Finding this equilibrium ensures that our GenAI solutions are not just efficient but also effective in real-world scenarios. Anthropic Claude emerges as a frontrunner in the realm of LLMs, distinguished by its exceptional capabilities and groundbreaking advancements. Boasting unparalleled scale, efficiency, and adaptability, Anthropic Claude represents a pinnacle of GenAI innovation. Its ability to seamlessly integrate with Amazon Bedrock amplifies its prowess, providing users with an unparalleled ecosystem for knowledge integration and generation. With state-of-the-art architecture and a commitment to pushing the boundaries of what's possible, Anthropic Claude stands as a testament to the transformative potential of GenAI in shaping the future of technology and beyond. Prompt engineering holds the key to maximizing the potential of applications. Crafting prompts that consider persona, context, tags, and positioning can significantly enhance the model's ability to comprehend and generate relevant responses. This strategic approach empowers AI systems to better serve their intended purposes. In the realm of GenAI architectures, a robust knowledge base is essential. Amazon Bedrock stands out as a key component for architectures like RAG (Retrieval-Augmented Generation), providing a solid foundation to fuel intelligent decision-making and content generation. As we march towards AI integration, bridging the gap with legacy applications is crucial. Agents equipped with open APIs offer a seamless way to facilitate communication and interoperability, ensuring a smooth transition into the AI-driven future. Let's continue to explore, innovate, and collaborate as we navigate the exciting frontier of GenAI together! #aws #LanguageModels #Innovation #TechTrends #AIIntegration #webdevcon #claude #bedrock

68

While this is a hardware story, there are tons of parallels in #productengineering for #softwareengineering #product. A quick read too, Thanks to https://lnkd.in/gSBhy5aU

3

Focusing on the fundamentals of "garbage in, garbage out," this new update will open up the potential for public data to reduce illicit seafood in supply chains. Special thanks to our advisor, Roberta Carraro, for introducing some best practices of the Herokai. Shipping weekly makes it OK to be smoll, galvanizes teams and helps customers grow with you!

7

1 Comment

Substack recently moved their ML and AI pipelines over to Modal, from AWS SageMaker. We've had a lot of fun working with them and we just wrote about it on our blog. “Using SageMaker just felt like a very convoluted process. It felt very removed from a normal engineering workflow. With Modal, it’s a lot faster. You don’t feel like you have to wait for notebooks to spin up or remember to turn them off or any of that type of stuff.” – quote from Mike Cohen, Head of AI & ML Engineering at Substack In the process, we actually made several improvements to Modal's system latency, since Substack's use case is very latency-sensitive. These optimizations now benefit all Modal users. Take a look: https://lnkd.in/e8yMCiti

185

5 Comments

One of my favorite Maven Showcase projects this week is the LA County Crime Mapping Tool by Tai C. Here's what I love about this one: 👉 The analysis is directly tied to a clear and specific objective – targeting locations in LA County to launch decriminalization ad campaigns. 👉 The output is simple (a single map visual) but provides the exact information that stakeholders need 👉 The interactivity gives users the ability to drill up or down to explore trends at different levels of granularity My only feedback might be to include some specific takeaways and recommendations in the project description itself – such a great way to showcase those analytical thinking skills 💪 Link to the project details in the comments – check it out and leave Tai some love!

44

1 Comment

On May 16, I get the privelege of facilitating a discussion with some AI experts and enthusiasts in our webinar, "The C-Suite’s Guide to Generative AI". I hope #DeltekProjectNation will join me, #AI Strategist Walter Pasquarelli and Lora Barclay from Waterstons for this discussion on valuable AI insights and practical strategies. Don’t miss it. 💻 https://bit.ly/49Ns1BD

16

1 Comment

Exciting work from Meta and World Resources Institute who have used AI to create the most detailed map ever made of trees across the planet. This allows anyone to query the height of forest canopies or even individual trees (when they stand isolated enough). The data is now freely available. As a fun test, I used the provided browser (https://lnkd.in/ghxQFmxt) to find one of my favorite neighborhood trees - the giant sequoia in Seattle's Leschi Park. The data accurately shows this tree towering above the others in the park (screenshots in comments below). The team used AI at Meta's DINOv2 visual transformer architecture in order to infer the height of trees given just a satellite or aerial color image as input. This is a very similar technique to 3D photos that you may have seen in your Facebook feed (https://lnkd.in/gXCygk5H), where the depth of an image is computed via AI in order give a 3D effect. I'm looking forward to this work continuing and becoming an important way in which the world measures, reports, and verifies the carbon stored in trees. As industries (from big tech to airlines) purchase biosphere carbon credits to offset emissions, it's critical that they have the best data possible. Improved imagery and AI offer a path to getting there. More info in the project video below.

51

9 Comments

My new standard attack is evolving to: "Why are you anthropomorphizing LLMs? Would you anthropomorphize 7zip or JPEG?" I'm reposting this article to help reframe the discussion. LLMs are, in fact, super lossy compressors. That's it. That's all they are. They don't care about your feelings because super lossy compressors don't have feelings and they don't have care. Stochastic parrot, or my new improved knockoff, stochastic pigeon, do not get to the heart of the problem. A compression / decompression algorithm is not a sentient being. End of story. Argument over. #WrittenByMe https://lnkd.in/gGaqyRMg

5

12 Comments

🚀 Excited to share my recent interview with Harry Guinness for LeadDev on Retrieval Augmented Generation (RAG) and its potential to tackle generative AI hallucinations. In the article, we discuss: How RAG enhances large language models with up-to-date data. Benefits in reducing hallucinations and improving AI reliability. Practical applications and cost considerations. Challenges in deploying RAG pipelines. Proud to highlight how RAG can help establish "ground truth" and build traceable citations, enhancing AI transparency and trustworthiness. Read the full article here: https://lnkd.in/gjarDhBX #AI #GenerativeAI #RAG #MachineLearning #ArtificialIntelligence #LeadDev #TechInnovation

9

The topic of #AI comes up in one way or another in practically all of my meetings nowadays. It's the buzziest of buzzwords. But those conversations often fall short of giving real actionable AI ideas that apply to me or my role. But...if you're a #marketer or #EventPlanner looking to make the most of AI, check out all the topics (and actionable tips) that Cvent has compiled for you. 👇 https://okt.to/WhjOt0

4

This is a great short article on the current shortcomings of LLM AI and some critique of the discourse around it. Always nice to see a shout out to my favorite comparison, "the Chinese Room" thought experiment.

6

I love the insights. We're starting to see a new landscape of enterprise search in data. And to your point, I think today's search is even more crazily limited in data / lakehouses (like, come on, regex matches are still the only supported SQL search query in year 2024)

13

Project update! My application tracking extension is now doing fuzzier matching and injecting new information into LinkedIn (and other sites) to keep me from doubling up on applications. Next up: local storage and bulk import/export operations...

12

2 Comments

How long are you allowed to store user testing/interview recordings, and how does that impact your work? I’m collecting experiences and opinions on data retention policies for user research as I prepare for my #DWAI2024 demo.  So far I’ve heard about vastly different experiences from folks in different companies around the US, and across the world: 🚫 Teams that are not allowed to record ANY research 🎥 Teams with NO restrictions on recording at all ✋ Teams that have suddenly gone from very few restrictions to strict 90-day retention policies I’ve also heard a range of attitudes about restrictions on recordings, from indifference to significant concern. If you work in UX, product, design, or marketing, I’d love to hear what your experiences have been with this. Feel free to respond below or DM me!

5

How can you talk a large language model into NOT hallucinating on your data? Thanks to Geoff Livingston and Greg Verdino for a great podcast to explore that question and lots more. Get the whole pod here: https://lnkd.in/gFPkQHAy

54

I am so honored to participate in an evening of women leadership and data with some amazing lady leaders from various industries! I am looking forward to an evening of networking and learning from each other! #WomenInData #Empowerment #Networking #DataAnalytics

20