Maarten Ectors’ Post

View profile for Maarten Ectors, graphic

Multi-Award Winning Exponential Growth Innovator | NextTech fCTO/fCPO/fCRO | AI, Web3, IoT, Spatial, GreenTech | Helping you Launch and Grow Exponential NextTech Ventures

The next big thing in AI after LLMs are VLMs. Large Language Models allow computers to chat with humans. However using images or videos in those chats is still resulting in bad outcomes. Often when ChatGPT is asked to generate text inside images, there are spelling mistakes. This is where VLMs or Visual Language Models come in. They should be able to describe what is in an image or a video in words which afterwards can be used in an LLM. So imagine a VLM security camera which explains in text that there is a long queue in front of the two open shopping checkouts, employee X just served three lattes while employee y has been talking to employee z, an unknown person has just broken the window, a fox got into the chicken stables, a house fire started,... If VLMs can translate what happens in the world around us then robots will be able to interact a lot easier with it. Tesla self-driving software is already using VLMs to describe what is happening around it, e.g. stop sign at 50m. What do you think the future of VLMs will be? https://lnkd.in/eGsj9qb2

Meta Introduces Vision Language Models, Shows Superior Performance Over Traditional CNNs

Meta Introduces Vision Language Models, Shows Superior Performance Over Traditional CNNs

http://analyticsindiamag.com

Alexander Alten

Co-creator Apache Wayang | Tech Lead | Data Architectures | Author

1mo

SLM's, since the LLM doesn't help for personal requests. SLM's are the true winner, real personal assistants, tuning themselves to the individuals environments. I think the mobile phones and their respective voice system will make the cut. They have the reach, the power to learn and ask the swarm via FL, organizing and improvise.

Peter Cranstone

CEO@3PMobile l Reimagining Digital Engagement l Low-cost Growth Engine for Web-based Businesses l Harnessing the Power of Digital Ecosystems through Consumer Choice.

1mo

I have to smile - VLMs sound like LLMs except piled higher and deeper (Ph.D.)

See more comments

To view or add a comment, sign in

Explore topics