Ivan Bogatyy

With 3 days to go to the largest gathering of investors, innovators, and entrepreneurs in Southern California, our next session at TechCon SoCal will explore AI use cases across industries and sectors. Moderated by Misti Cain, Investment Manager at Techstars, the panel includes Badri Raghavan, VP of Data Science and AI/ML at ResMed, SK Sharma, Chief AI Officer at Universal Music Group, Ramsey Pryor, Head of Global Sales at Rumi.ai, and Lena Skliarova-Mordvinova, CTO and Co-Founder of Good Face Project. These experts will discuss how AI is transforming their respective fields. In today's landscape, AI is revolutionizing industries from healthcare to entertainment and beauty. The panelists will share insights on AI-driven personalized healthcare, music creation and distribution, sales optimization, and cosmetic product development. By leveraging AI, businesses are enhancing precision, improving customer experiences, and unlocking new revenue streams, showcasing AI's transformative power. Reflecting on the past, AI has evolved from theoretical research to practical applications. Initially limited to academic settings, advancements in technology have brought AI into the mainstream. Past TechCon SoCal sessions may have focused on AI's potential, but now the focus is on real-world implementations and success stories, highlighting AI's role as a key driver of innovation across various sectors. Don't miss this transformative AI session at TechCon SoCal—join us and get ready for industry insights! OPEN Silicon Valley TIE San Diego Interlock Capital Orange County OC Startup Council The Brink SBDC EvoNexus UC San Diego SDSU College of Sciences Pismo Ventures Cross Ocean Ventures Startup Steroid - Deal Flow | SPVs | Demo Day Platform Alliance for Southern California Innovation Spark Growth Ventures SoCal Startup Day Reflect Ventures Wharton Alumni Angels NuFund Venture Group Flok Labx Ventures Mercury Pankaj Kedia Serhat Pala Neal Bloom David Saxton Cheryl K Goodman Ellen M. Chang Niraj Desai Akash Pai Sufyan Subzwari Jonah Peake 🐺 Ella Napata Stephen Silver Fred Grier Scott Fox 🦊 Atlás Blake Eric Eide #TechconSoCal2024 #IndustryExperts #TechConference #InnovationSummit #TechTalks #startupsteroid #innovations #technologynews #networking #artificialintelligence #digitalhealth #robotics #startups #AngelInvestors #vc #SandDiegoStartup #founders #consumertech

11

2 Comments

Andrew Ling from Groq is hosting an insightful session on "Extending PyTorch for Custom Compiler Targets." Groq has pioneered the world’s first LPUs (Language Processing Unit), designed for LLM inference and deep learning acceleration. In this session, Andrew will delve into: 🔹 Maximizing inference efficiency for custom hardware targets by annotating PyTorch code with arbitrary information. 🔹 How to use PyTorch model annotations to adjust workload mapping to accelerators while preserving the semantics of the PyTorch inference graph. 🔹 A generalizable technique to inject and recover metadata in PyTorch graphs at different granularities, enabling custom data types and precision information through PyTorch into the Groq compiler. This talk is a must-attend for anyone pushing the boundaries of PyTorch and exploring custom hardware acceleration. See the full abstract here: https://lnkd.in/gwSRjPjz #MachineLearning #PyTorch #groq #hardware Toronto Machine Learning Society (TMLS)

39

LLMs can take gigabytes of memory to store, which limits what can be run on consumer hardware. But quantization can dramatically compress models, making a wider selection of models available to developers. You can often reduce model size by 4x or more while maintaining reasonable performance. In our new short course Quantization Fundamentals taught by Hugging Face's Younes Belkada and Marc Sun, you'll: - Learn how to quantize nearly any open source model - Use int8 and bfloat16 (Brain float 16) data types to load and run LLMs using PyTorch and the Hugging Face Transformers library - Dive into the technical details of linear quantization to map 32-bit floats to 8-bit integers As models get bigger and bigger, quantization becomes more important for making models practical and accessible. Please check out the course here: https://lnkd.in/g66yNW8W

10,609

143 Comments

Savala stands at the forefront of sustainable commodity trading, pioneering a new era of ethical investment in the graphite market. By harnessing the power of blockchain technology, Savala ensures unparalleled transparency and accountability throughout the trading process. In a world where transparency is paramount, Savala leads by example. Each transaction on their platform is meticulously verified, promoting ethical trading practices and setting a new standard for integrity in the industry. This commitment is especially critical in commodities like graphite, a key component of lithium-ion batteries. But Savala goes beyond mere transactions – Savala cultivate connections. By facilitating direct relationships between mines and buyers, Savala empowers sustainable practices and rewards eco-conscious investments. Their platform serves as a catalyst for positive change, fostering a community where environmental responsibility is celebrated. In the realm of commodity trading, information is power. Savala provides real-time data and insights, enabling fair pricing and supporting sustainable mining endeavors. It's not just about profits; it's about building a future where environmental stewardship and economic prosperity go hand in hand. Ready to join the movement? Invest with Savala and support commodities that power our future while championing a planet-friendly approach. Together, let's be the change we wish to see. ✅ Follow Savala for more information #SustainabilityInTrading #BlockchainForGood #Savala

5

I am now open-sourcing my full winning solution for the Optiver competition, which developed a model capable of predicting the closing price movements of hundreds of Nasdaq-listed stocks. The model became the prize-winner among over 4,000+ quant and data scientist teams. Optiver is a trading firm with over $65 billion USD in assets under management and over $1.7 Billion in revenue. The model is a blend of a tree-based model, a feedforward, and a 1D convolutional neural network, presented in the training and inference folders. Please let me know if you have any questions. #deeplearning #ML #quant #AI Kaggle

1,571

45 Comments

Scoop w/Riddhi Kanetkar: Holistic, a new AI startup founded by ex-Deepmind scientists, is set to raise $200 million in initial funding, sources told Business Insider. The Paris-based company is bringing funding in across a large equity financing and a convertible note, with a bevy of leading investors taking part. It’s understood that Accel are leading the round, with participation from AI software giant UiPath and BPI France. Following on from the success of Mistral, France is now a hotbed for businesses building foundational models and LLMs in AI. You can read more about the deal’s structure and valuation here: https://lnkd.in/ernn5ze4 #vc #ai #startups #tech #LLMs #funding #capital

45

5 Comments

I'm so excited to share that TLM v1.0 has been officially launched! 🎉 TLM is where Data Curation 🤝 LLMs. Thanks to the wonderful efforts of the Cleanlab team, we've created an awesome product that can: - Improve LLM response reliability - Quantify the trustworthiness of responses from any LLM LLMs will always have some hallucinations, but by providing a trustworthiness score with every output, Cleanlab TLM lets you identify when the LLM is hallucinating to enable reliable deployment of LLM-based applications. Read more about it in our research blog: https://lnkd.in/dTZUVTmG Check it out for yourself here: https://tlm.cleanlab.ai/

31

1 Comment

OpenPipe Introduces a New Family of ‘Mixture of Agents’ MoA Models Optimized for Generating Synthetic Training Data: Outperform GPT-4 at 1/25th the Cost Quick read: https://lnkd.in/gfxdiDT5 OpenPipe’s MoA models have excelled in rigorous benchmarking tests, achieving notable scores on LMSYS’s Arena Hard Auto and AlpacaEval 2.0. The MoA model scored 84.8 on Arena Hard Auto and 68.4 on AlpacaEval 2.0, indicating its superior performance in generating high-quality synthetic data. These benchmarks are critical as they represent challenging user queries that test the robustness and adaptability of AI models. The MoA model has been benchmarked against various GPT-4 variants in real-world scenarios. Results showed that OpenPipe’s MoA model was preferred over GPT-4 in 59.5% of the tasks evaluated by Claude 3 Opus. This is a significant achievement, highlighting the model’s effectiveness and practical applicability in diverse tasks encountered by OpenPipe’s customers...... OpenPipe

43

3 Comments

Wondering what top research got published this last week? 📑🤖🤔 So essentially, 1. "AI Model StyleFeatureEditor gives highly detailed image editing results!" 2. "We have a Universal Quantum Chemistry dataset to find new drugs now!" 3. "Stylebreeder clusters diverse art styles from 6.8M images on Artbreeder, enhancing personalized AI-driven artistic expression." 4. "LongRAG addresses the limitations of traditional RAG by using long retrieval units" 5. "Language model tuned to avoid toxic language in English will also avoid it in other languages!" 6. "YOUDREAM Auto-Generates Perfect 3D Creatures from Text!" From cutting-edge research on Artificial Intelligence to advances in Machine Learning, Computer Vision, and Neuroscience, we distill the latest technical papers into bite-sized insights you can apply today! For more subscribe at https://lnkd.in/gJ_wJxAC

21

1 Comment

A New Machine Learning Research from UCLA Uncovers Unexpected Irregularities and Non-Smoothness in LLMs’ In-Context Decision Boundaries https://lnkd.in/dbbCCVGJ Practical Solutions and Value of In-Context Learning in Large Language Models (LLMs) Understanding In-Context Learning Recent language models like GPT-3+ have shown remarkable performance improvements by predicting the next word in a sequence. In-context learning allows the model to learn tasks without explicit training, and factors like prompts, model size, and order of examples significantly impact results. Exploring Methods of In-Context Learning This paper explores three methods of in-context learning in transformers and large language models (LLMs) through binary classification tasks (BCTs) under varying conditions. It aims to link in-context learning with gradient descent, understand its practical application in LLMs, and utilize MetaICL for enabling in-context learning. Research Findings and Experiments Experiments focused on evaluating pre-trained LLMs’ performance on BCTs, understanding the influence of different factors on decision boundaries, and improving their smoothness. The decision boundary of LLMs was explored for classification tasks by prompting them with in-context examples, and the results demonstrated the non-smooth nature of these boundaries. Implications and Future Insights Despite high test accuracy, the decision boundaries of LLMs were found to be non-smooth, and factors affecting this were identified through experiments. Fine-tuning and adaptive sampling methods were explored to improve the smoothness of the boundaries, providing new insights into the mechanics of in-context learning and pathways for research and optimization. AI Solutions for Business Evolution Evolve your company with AI to stay competitive and redefine your way of work. Identify automation opportunities, define KPIs, select AI solutions, and implement gradually to leverage the benefits of AI. For AI KPI management advice and continuous insights into leveraging AI, connect with us at hello@itinai.com and stay tuned on our Telegram t.me/itinainews or Twitter @itinaicom. Discover how AI can redefine your sales processes and customer engagement. Explore solutions at itinai.com. List of Useful Links: AI Lab in Telegram @itinai – free consultation Twitter – @itinaicom #InContextLearning #LLMs #MachineLearning #AIResearch #BusinessEvolution#artificialintelligence #ai #machinelearning #technology #datascience #python #deeplearning #programming #tech #robotics #innovation #bigdata #coding #iot #computerscience #data #dataanalytics #business #engineering #robot #datascientist #art #software #automation #analytics #ml #pythonprogramming #programmer #digitaltransformation #developer

1

📣 With this workflow, we enable our users to effectively A/B test between different versions of a prompt and measure the respective accuracy. This empowers developers to make data driven decisions about choosing a prompt when building with LLMs.

6

Loved chatting all things investing in the next wave of AI with Prateek Joshi on the Infinite Machine Learning Podcast! At March Capital we've been investing in AI for a decade and view it as a fundamental technology driving enterprise value. Check out the episode link below to hear us chat more about metrics that matter, finding the signal through the noise in an AI crazed world, and AI's future impact on both business and our daily lives!

24

1 Comment

Look forward to speaking at the Enterprise Generative AI Summit this week along side esteemed panelists Dr. Yvonne Lutsch and Mike Shirazi on the importance of data relevancy and data privacy considerations in next wave of LLM investments. #dataprivacy #GenAI #LLM https://lnkd.in/efj2egAG

81

As someone who follows the world of AI closely, I was excited to read about Sandeep Nailwal's new venture, Sentient. They just raised a whopping $85 million to build an open-source, decentralized AI platform. That's a big deal! 𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗰𝗮𝘂𝗴𝗵𝘁 𝗺𝘆 𝗲𝘆𝗲: 𝗢𝗽𝗲𝗻𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝗧𝗿𝗮𝗻𝘀𝗽𝗮𝗿𝗲𝗻𝗰𝘆:Sentient is built on blockchain, which means everything is open and verifiable. No more shady algorithms! This is a big contrast to closed-source projects like some from Google and Meta. 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 𝗣𝗼𝘄𝗲𝗿:Sentient is all about the community. Developers will be rewarded for their contributions, which is a major shift from the current model where big companies reap all the benefits. 𝗙𝗮𝗶𝗿𝗻𝗲𝘀𝘀 𝗳𝗼𝗿 𝗔𝗹𝗹:Nailwal believes AI development shouldn't be controlled by a few giants. Sentient aims to level the playing field and create an "AI economy for all." 𝗡𝗼𝘄, 𝘀𝗼𝗺𝗲 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀 𝗿𝗲𝗺𝗮𝗶𝗻: 𝗖𝗮𝗻 𝘁𝗵𝗲𝘆 𝗰𝗼𝗺𝗽𝗲𝘁𝗲 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗯𝗶𝗴 𝗴𝘂𝘆𝘀?OpenAI, Google AI—these are established players with massive resources. Can a startup like Sentient truly 𝗜𝘀 𝗮 𝘁𝗼𝗸𝗲𝗻 𝗰𝗼𝗺𝗶𝗻𝗴? They haven't announced a token yet, but it sounds like they might in the future. This could be a game-changer for incentivizing the community. Overall, Sentient is a breath of fresh air in the AI space. Their focus on openness, fairness and community engagement is exciting. While they have a tough road ahead, I am definitely keeping an eye on their progress. This could be the start of something big! 𝗪𝗵𝗮𝘁 𝗱𝗼 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸 𝗮𝗯𝗼𝘂𝘁 𝗦𝗲𝗻𝘁𝗶𝗲𝗻𝘁? 𝗟𝗲𝘁'𝘀 𝗱𝗶𝘀𝗰𝘂𝘀𝘀 𝗶𝗻 𝘁𝗵𝗲 𝗰𝗼𝗺𝗺𝗲𝗻𝘁𝘀! https://lnkd.in/d_U5syCi #ArtificialIntelligence #AI #MachineLearning #FutureofAI #AIRevolution #OpenAI #DecentralizedAI #BlockchainAI #CommunityDrivenAI #FairAI

8

1 Comment

this twitter thread from Eric Paley on the 2 laws of #startup physics is truly epic. Most original and well-written piece I've read in a while. https://lnkd.in/gD3BSKFz I seriously think all of us should print this out and hang it in our offices. Below are my top 10 of Eric's 55 tweets. # 10 ends with a killer quote to be internalized and committed to memory and practice. 1) The positive compounding formula Experiment with small amounts of capital until a formula is found that generates a surplus of intrinsic value on each dollar invested. Then, and only then, should you deploy capital in an attempt to scale that specific formula. 2) Capital can scale intrinsic value rapidly if you have an engine that allows it to turn $1 into $5 of value. If it has an engine that turns $5 into $1 (or even $1 into $0.99), capital will ultimately compound the negative value formula to zero. 3) Certainly, many companies create intrinsic value ahead of measurable economics. Still, it can be dangerous to believe you can spend millions with no source of true validation – and venture capital is a poor source of validation. 4) All positive systems compound at diminishing rates of return Sooner or later, the return on each dollar invested will shift to negative. If you’re unaware of the point at which compounding goes into the red, you start compounding negative value. 5) Capital creates inertia The faster you spend capital, the harder it is to stop. As startups scale, stopping a dollar of spending is harder than deciding to spend it in the first place. 6) It’s easy to rationalize diminishing returns Founders launch companies because they believe in their vision. VCs invest because they want extraordinary results. Diminishing returns are easy to ignore and truth becomes harder and harder to engage. 7) The bill comes due suddenly A startup can compound negative value for months (even years), but the scale of the loss may only become apparent at an inflection point, typically when trying to raise capital - particularly when the capital markets have tightened. 8) Most VCs want growth so badly that they’ve come to demand it regardless of the cost. This obsession is a short-term optimization with huge long-term implications for the startup, its founders, employees, and, ultimately, investors. 9) Entrepreneurs respond to this market demand by spending poorly on activities they know drive vanity metrics that aren’t sustainable to ensure their revenue continues to grow to meet their short-term goal of raising more money. 10) Instead, focus on low-cost experiments until you discover how to create positive value formulas that justify increased spending. Focus on learning fast, not burning fast.

35

9 Comments

pumped to announce sacra pipeline, which syncs sacra research and data directly into your database or data warehouse—so you can work with it natively using your existing tooling. our mission at sacra is to get our research & data on growth and pre-ipo tech companies into the hands of every platform and fund that needs it. sacra pipeline makes that stupidly easy by piping our data to you from day 1. just connect your database (postgres, mysql, mongodb) or data warehouse (snowflake, redshift, databricks) and your product and data teams can start cranking immediately. to check it out, see our launch vid below and email us at founders@sacra.com to get started 👇 🔌 for platforms, this means exposing sacra data in customer-facing features without building data pipelines 🧑‍💻 for funds, it means analyzing companies in sql and creating prospect lists without switching tools. 🔁 for ai developers, it means ingesting and continuously syncing data for RAG as we add to and update our research ultimately, our goal is "easy in, easy out" data. stupidly easy to sign up, get started, and get value - without barriers or commitments. major thanks to Danny Tharma who made it happen here at sacra! to get started with sacra pipeline, email us at founders@sacra.com. 🚀

26

5 Comments

Excerpt from an Angel Squad memo that we invested in recently: “[Company] is led by a team of scientists and engineers with PhD-level expertise in Machine Learning, Engineering Simulations, and Computational Materials Science. The team is headed by CEO, [Name], who previously led AI development at two early-stage startups. [Name] is joined by CTO, [Name], who has deep experience in machine learning and materials science, and Chief Science Officer, [Name], a doctoral researcher at the [School]. “ Investing at the earliest stages of value creation often means taking big bets on a founding team. At Angel Squad and Hustle Fund, we love founders with deep expertise and super domain knowledge, particularly in deep tech verticals.

28

3 Comments

With 25 days to go to the largest gathering of investors, innovators, and entrepreneurs in Southern California. Meet Murali Joshi, Partner at ICONIQ Growth, our next speaker at TechCon SoCal. Murali Joshi, a Partner at ICONIQ Growth, brings nearly 6 years of experience to the forefront of venture capital. With a keen eye for promising opportunities, Murali has invested over $2.2B across 19 portfolio companies, including Datadog, Procore, and Relativity. Currently, he serves on the board of directors for notable firms such as 1Password, Drata, and Orca Security, driving growth and innovation in the tech landscape. Prior to his role at ICONIQ Growth, Murali honed his expertise in investment banking at Goldman Sachs, where he worked closely with leading technology companies like Twilio, Instacart, and Zoom. His entrepreneurial spirit led him to co-found a software analytics startup before joining the venture capital realm. Recognized on Forbes' 30 Under 30 for Venture Capital, Murali's contributions to the industry continue to shape the future of tech investment. Join us on May 24th, 2024, as Murali shares insights on investment strategies shaping the future of technology! 🚀 OPEN Silicon Valley TIE San Diego Interlock Capital Orange County OC Startup Council The Brink SBDC EvoNexus UC San Diego SDSU College of Sciences Pismo Ventures Cross Ocean Ventures Startup Steroid - Deal Flow | SPVs | Demo Day Platform Alliance for Southern California Innovation Spark Growth Ventures SoCal Startup Day Reflect Ventures Wharton Alumni Angels NuFund Venture Group Flok Labx Ventures Mercury Pankaj Kedia Serhat Pala Neal Bloom David Saxton Cheryl K Goodman Ellen M. Chang Niraj Desai Akash Pai Sufyan Subzwari Jonah Peake 🐺 Ella Napata Stephen Silver Fred Grier Scott Fox 🦊 Atlás Blake Eric Eide ICONIQ Capital #TechconSoCal2024 #IndustryExperts #TechConference #InnovationSummit #TechTalks #startupsteroid #innovations #technologynews #networking #artificialintelligence #digitalhealth #robotics #startups #AngelInvestors #vc #SandDiegoStartup #founders #consumertech

11