Microsoft just dropped OmniParser V2, a screenshot parser for web automation!
You can now turn any LLM of your choice, DeepSeek R1, GPT-4o/o1, or Qwen 2.5VL, into a Computer Use Agent.
This is 60% faster than V1
100% Open Source.
New 3h31m video on YouTube:
"Deep Dive into LLMs like ChatGPT"
This is a general audience deep dive into the Large Language Model (LLM) AI technology that powers ChatGPT and related products. It is covers the full training stack of how the models are developed, along with mental models of how to think about their "psychology", and how to get the best use them in practical applications.
We cover all the major stages:
1. pretraining: data, tokenization, Transformer neural network I/O and internals, inference, GPT-2 training example, Llama 3.1 base inference examples
2. supervised finetuning: conversations data, "LLM Psychology": hallucinations, tool use, knowledge/working memory, knowledge of self, models need tokens to think, spelling, jagged intelligence
3. reinforcement learning: practice makes perfect, DeepSeek-R1, AlphaGo, RLHF.
I designed this video for the "general audience" track of my videos, which I believe are accessible to most people, even without technical background. It should give you an intuitive understanding of the full training pipeline of LLMs like ChatGPT, with many examples along the way, and maybe some ways of thinking around current capabilities, where we are, and what's coming.
(Also, I have one "Intro to LLMs" video already from ~year ago, but that is just a re-recording of a random talk, so I wanted to loop around and do a lot more comprehensive version of this topic. They can still be combined, as the talk goes a lot deeper into other topics, e.g. LLM OS and LLM Security)
Hope it's fun & useful!
https://t.co/75mXcUBI8L
This is wild.
Anthropic just dropped Computer Use and Claude 3.5 Sonnet & Haiku
And it will completely change the AI agent game.
6 wild examples:
1. Computer Use for Coding
Big News! 🚀 CitySwift's funding round hits €7 million, steering towards success in transportation optimization!
Discover how @CitySwift 's AI-powered analytics are reshaping public transport.🚌
Congratulations @BrianO_Rourke Alan Farrelly and team👏 https://t.co/dgEdqrt7l2
We’re delighted to announce the closure of €7mn funding, led by Gresham House Ventures🚀
The investment will be used to accelerate development, providing both authorities and operators with intelligent data solutions to future-proof transport networks 🚌
https://t.co/SO2OHeAZP7
Delighted to announce we're partnering with @Translink_NI to enhance Belfast's bus network.
CitySwift will enable Translink to put data at the core of improving operations and collaborating on network planning.
https://t.co/eCfc8mCwRc
#PublicTransport#DecisionIntelligence
We’re excited to announce that Metroline is partnering with CitySwift to speed up data analysis and power reliable bus schedules in London.
Learn more about how Metroline are using data to improve the passenger experience: https://t.co/4oajaguKeR
#DecisionIntelligence
Sligo surfer beats current world champion Filipe Toledo 🇮🇪
The Irish surf team are in El Salvador this week competing in the world surfing games. Congrats to Ireland’s Gearoid McDaid who took on the world champ and won in his last heat.
@gearoidmcdaid 👊
INTRODUCING: CITYSWIFT SIMULATIONS
We are thrilled to announce our new game-changing scenario planning and predictions tool: CitySwift Simulations! 🎉
Read more about our latest innovation in the most recent issue of @passtrans magazine 📰 https://t.co/4YpUa1NM5o
E111 ‼️
BIG BREAKDOWN on $MSFT investing $10B into @openai
-- unique deal terms
-- $META & $GOOG asleep at the wheel
-- what startups can be built
-- how to make the business model work
+ much more
🔊: https://t.co/vXVbFxRFug
📺: https://t.co/UVHzQJhBCY
If you could come up with a new name for chatGPT, what name would you choose?!
Retweeting this message is appreciated, I want to hear as many names as possible.
Best name gets eternal fame! 🏆
#chatgpt