Cosmos Darwin

Meta is facing BIG restructuring toward innovation and growth. CTO, Andrew “Boz” Bosworth, has unveiled the most extensive restructuring of Reality Labs to date, splitting our hardware division into two focused groups: Metaverse and Wearables. 🔷 Metaverse: This team now includes our cutting-edge Quest headset line and is dedicated to pushing the boundaries of immersive experiences. 🔷 Wearables: Encompassing all other hardware efforts, including the innovative smart glasses collaboration with Ray-Ban. This strategic move aims to streamline and sharpen our focus on delivering groundbreaking products in these two domains.  #Meta #Innovation #Wearables #Metaverse #TechTransformation https://lnkd.in/gQ5RdydW

5

I've spent multiple semesters in the SAC and learned a lot about my own investment thesis, learning the mechanics of investing and meeting fascinating startups! If you're on the fence about angel investing, this is a good starting point

11

With Llama 3.1, Meta takes another step towards becoming the “Windows of AI”. Meta releases their Q2 earnings this week, the first since they dropped the first private-grade open source foundational model , a huge step towards this strategy. Investors will be asking what Meta's long-term vision for AI is, and evaluating how likely Meta is to eventually "cash in" on their heavy investments to produce foundational models. Some early thoughts on how to think about Meta's AI strategy:

13

Thats the exact explanation I gave to a group of investors last week. (… it’s just Precision/Recall, attributes vectorization as “embedding”, K-Nearest Neighbor, User Search Intent, DAG, ML algorithms, etc all over again …). LOL and they asked what’s new this time around. And I was like not much, ML is still the backbone, and AI is the inference and actions based on the model. It works better when there is a feedback loop in a closed system. But the current state of the art like ChatGPT is based on scraping the Internet (garbage in, garbage out). Kinda like early search engines. What is a better analogy is that NVIDIA/Snowflake etc are selling tools, aka pickaxes to prospectors. The gold is up to you to find. And the opportunity right now is to monetize industry specific (healthcare, travel, financial services) clean data to build models with (aka a treasure map for the prospectors). From an enterprise perspective, what’s out there today aka generative AI isn’t very useful. Too much hype. To build useful engines will require an effort similar to search revolution 2 decades ago. From

16

3 Comments

Human in the loop with AI: Best-in-class LLMs are now at ~80th percentile in output quality. That is excellent for first drafts and being better than average, but still not great for delivering a final product of specialized, expert level work. #LLMs #AI #GenAI #futureofwork https://lnkd.in/gpe6BJjx

12

We've learned a lot in the last year building #AI at Microsoft on how to do safely and responsibly. Here's some of the themes we've found while doing this work. https://lnkd.in/g8vy2kur

248

3 Comments

Last night I had the fortune of representing LaunchDarkly at Sourcegraph's CodeAI Summit NYC. An important theme I took from last night's event? Context. Models like specifics. Computers can't read our brains or know what we want to achieve. Technology is kinda useless without the right context. As we move forward in this increasingly AI-generated world (for better and for worse) let us not forget the importance of adding context to the equation. Whether that's providing context to a dataset through how it was collected, sharing context of your repo to better aid a code assist tool, or simply providing context behind the why of our decisions. It's an interesting (and important) place to keep in our minds. #AI #MachineLearning #DeveloperAdvocacy #DeveloperEvents

14

1 Comment

Google has begun rolling out private access to its Veo and Imagen 3 generative #AI models. Starting Dec 4th, customers of the company’s Vertex AI Google Cloud package can begin using #Veo to generate videos from text prompts and images. Then, as of next week, Google will make Imagen 3, its latest #texttoimage framework, available to those same users. #designinspiration #GenAI #GenerativeAI #feedly via Engadget

1

I am excited to share the latest blog, co-authored by myself, Junhwi Kim and Vyas Sekar, detailing how we managed to successfully localized and improved LLM by combining Fine-Tuning and RAG. https://lnkd.in/gKDkgGPq

70

3 Comments

Building regulation around AI requires a conversation that supports both innovation and responsibility. I'm happy to see that bills are being drafted and vetoed—it's a process we go through to make things better. I just hope this is a trend that leads towards more bills being presented and more discourse being had in public about this important topic. At the same time, the tech industry needs to think responsibly about guardrails. It's why I'm optimistic about companies like meetsynthia.ai and Squirro who are building real-world solutions today to help organizations leverage Generative AI solutions today while maintaining responsible guardrails. Drew Rayman

3

1 Comment

🍜Anybody up for some ramen? I am excited to share that I am back in the garage, building again (it’s actually a house in the Berkeley Hills). We are taking the wraps off of Supercog AI, a new startup focused on solving Application Integration using LLM powered agents. My co-founder in this new venture is a great friend of mine (and a former co-founder partner), Alex Osborne. Alex and I have worked around the application integration space for a long time, and we are unrealistically excited about applying the GenAI stack to this problem. Not many people relish the task to get some bits to move from system A to system B. But it’s a critical job to be done, and that’s led to a lot of brittle scripts, a myriad of inflexible “no code” tools, and a lot of hand crafted SQL. But the power of the Large Language Model, trained on a huge corpus of information, offers a remarkable tool to solve this problem. The LLM you use today already knows the APIs of hundreds of popular systems. It knows the SQL dialects of every major database. It understands the semantics and data schemas of many popular SaaS systems. It understands much of the specific domain in which your business operates. I like to call the AI revolution “the last platform shift”. That may turn out to be hyperbole, but I definitely believe we are in the very early stages of seeing what this new stack can enable. I know we are hardly the first folks to claim that “amazing stuff is coming!”. But this is why Alex and I have decided to focus on a real and hard problem. If we can prove that LLMs can power a new way to solve this problem, it will be the first successful *new* approach in 15 years.. We will have much more to share about this new platform soon. In the meantime, if you’re interested in getting a peek - or even better if you have application or data integration tasks that you would like some help with - please reach out because we want to help. #genai #startups #backinthegarage

163

37 Comments

🚀 Unlocking the Power of Local LLMs with Ollama! 🚀 I've been diving into the world of local LLMs using Ollama, and I must say, the speed of the Meta Llama 3.2 1B parameter model on my HP Elitebook is nothing short of impressive. While it doesn't quite match the speed of online models, it's performing exceptionally well for building out my own local LLM. I've even been able to successfully integrated local knowledge with Retrieval Augmented Generation (RAG) to answer questions using my own data. 😮 Next stop: local Agent AI! There's a bit of a learning curve, but you can get started with: - Ollama: Ollama - Download the 'llama3.2:1b' model. - OpenWebUI: OpenWebUI to simplify the RAG process. Best of all, it's all free! 🤩

23

7 Comments

You've heard chain-of-thought reasoning, but have you heard of COCONUT reasoning? A new development from Meta and UC San Diego that I'm actually excited about: chain-of-continuous-thought (COCONUT). AI researchers are now pushing reasoning back into the latent space of the transformer rather than mediating it all through text, like standard CoT. I think this is a massive step forward from the extremely "meh" progress of GPT-o1. We're getting closer to how humans actually think (spoiler alert: you do not reason in language), which opens up new possibilities for learning, memory reconsolidation, and generalization. Is this a new architecture that could extend far enough to beat ARC-AGI? Paper here: https://lnkd.in/gKNSERKp

11

1 Comment

Attention SoCal engineering leaders! ELC is hosting two free events in November: 📍 Santa Monica / LA – Nov 14: Don’t miss an exclusive live interview with renowned speaker and author Michael Feathers. Get a preview of insights from his upcoming book, AI Assisted Programming, and get a clear picture of how AI can (and can’t) enhance your team’s software practices. This conversation will cover concrete strategies and will steer clear of the usual AI hyperbole. 📍 San Diego – Nov 21: Join us for a roundtable on Practical Measurement & Tracking of Engineering Metrics. Share ideas with peers and separate what truly drives results—and what incentivizes the wrong things. 🔗 Registration links are in the comments below. Attend to network, learn, and grow!

10

5 Comments

Really great from the frequently great Mariusz Ciesla "Let's face it: contrary to what chronically online LLM nerds say, nobody wants to learn to prompt. Building front-ends that get the user to their stated success without forcing them into the mess of prompt engineering is where I can see AI-based products going. The models get better, the interfaces should too.... your job is likely safe, and being skeptical about the model output is likely warranted, unless you're using a very well finetuned model" https://lnkd.in/g9rFsUA2

2

Don't fight against AI. Help it to get latest info. 👇 is the talk I gave last week regarding Real-Time #RAG in Zilliz's Unstructured Data and #LLM meetup. "Talk is cheap, show me the code" This blog shows the key data flow and code blocks. The full example is available at https://lnkd.in/gUj349hh How it works? Input: Stream stories and comments from #HackerNews API. Pre-process: Retrieve updates and filter for stories/comments via #Bytewax. Retrieve Content: Download the html and parse it into usable text. Thanks to the awesome #Unstructured.io. Vectorize: Create an embedding or list of embeddings for text using #HuggingFace #Transformers. Output: Write the JSON document with #embeddings and other fields to a local #Timeplus #Proton server, into the data streams. Filter/Route: Use streaming SQL to filter/transform JSON data and send output to a local #Kafka server. Forward: Use #Milvus plugin for #KafkaConnect to forward the data to #Zilliz Cloud. Query: Build a #streamlit web application to query the live data with the same embedding model.

18

1 Comment

When Google first launched Dialog Flow, everyone was excited about its bold promise: creating easy, seamless dialog flows using natural language. It aimed to make sophisticated conversational AI accessible to everyone, regardless of their technical expertise. The reality as we all know -- was a bit different. It's amazing to see how far we have come with LLMs. Once you get you used to designing conversations flows like this, it's hard to go back. #unSurvey's conversation flows are dynamic, context-aware and SUPER easy to design. #ConversationalAI #DialogFlow #unSurvey #AI #LLMs

13

1 Comment