Enterprise VC-engineer-company builder. Early investor in @databricks, @tubi and 6 other unicorns - @cohesity, @eightfold, @turing, @anyscale, @alation, @amperity, | GP@Foundation Capital
I’ve been in the AI trenches since 2009, and LLMs are certainly a game-changer. But they also seem to be a warm-up act for the main event—the next cycle of AI innovation, coming in the next 12-18 months. Here are 3 areas we’re looking at to fuel this cycle, where founders can make real AI magic: – Multimodality 📸 – Multiple agents 🤖🤖🤖 – Post-transformer architectures ⚒️ Blog in the comments ⬇️
Thank you for sharing your insights on the next cycle of AI innovation. I would like to humbly present NetAI, founded by ex-Google Data Scientist, as an example of this upcoming wave of innovation. NetAI leverages multiple agents with Graph Neural Networks (GNN) instead of Large Language Models (LLMs) in the realm of AIOps, and here’s why: Structural Relationships: LLMs capture the structural relationships between words, which is perfect for natural language processing but less suitable for computer networks and transport links. This limitation can make LLMs less effective in scenarios where understanding network topology and behavior is crucial. The approach is only as good as the data used to train the model, and similar to rule-based methods, it may face challenges in the long run. Graph Neural Networks: GNNs excel in capturing the structural relationships between nodes and edges, offering a model that closely mirrors real-world networks and transport links. This capability allows GNNs to provide deeper insights and more accurate modeling of network behaviors and anomalies, making them highly effective for AIOps applications. We believe GNN has more headroom for innovation than LLM for AIOPS
I feel like Deepmind and Turing are being slept on within this article. Tiktok disrupted/expanded the market utterly for content creation, influencers, and B2B. Until Instagram reels, YouTube shorts, Snapchat and others sought to copy paste their recipe. What wasn't regarded retroactively were the cooks in the kitchen working years on the TikTok recipe. Demis Hassabis since his world changing doctoral dissertation. Has been R&D cooking ever since he finished his PhD. Turing has expanded their kitchen with the top talent in the field of data science. Just as we saw with the UK. They sought to only talk about their chips and biscuits. Yet the majority of the food they ate/eat now. Comes from a country of dishes that were stewed hours over cow created fuel. Now they can't live without its cuisine, and it's influence has shaped Michelin star restaurants utterly. Yet, no one talk about how the long stewed recipe is the greatest. Only the folks who quickly put up something with awkward PMF, but can suddenly claim massive amounts of bubble based hype. The market leaders rarely lead or influence, but they do acquire those who do.
100% agree. At this point though, getting business stakeholders on board is key. The tech is either mature enough today or will be in the next 12-18 months to your point. I just do not like the term “agents”. It might be just semantics but I really prefer “resources” so another stakeholder conversation does not end with speculation about terminator robots.
Thanks for sharing the ideas and your thoughts! Love what Karan and the team are doing at Cartesia. Closely following. InteractiveAI and XAI in it’s next phase are going to be phenomenal and I am looking forward to it. GUI is going to be dead, interaction via voice and expressions is where it will head towards. I always say that we have been in a phase for a very long time where tech is static - it does exactly what we tell it do. But, that is about to change very soon! #AI #XAI #InteractiveAI
Bil Ahmed
2w
Isn’t the warm act for DI ‘Decsion Intelligence’ where we are able to make better decisions across all industries and use cases through cognitive mode? 😁
Enterprise VC-engineer-company builder. Early investor in @databricks, @tubi and 6 other unicorns - @cohesity, @eightfold, @turing, @anyscale, @alation, @amperity, | GP@Foundation Capital
3whttps://foundationcapital.com/beyond-llms-building-magic/