Log inSign up
Sam Witteveen
4,029 posts
user avatar
Sam Witteveen
@Sam_Witteveen
Co-founder @ Red Dragon AI, Google Developer Expert for Machine Learning/ Deep Learning,
Singapore / San Francisco
youtube.com/@samwitteveenai
Joined April 2008
1,725
Following
22.5K
Followers

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    Sam Witteveen
    @Sam_Witteveen
    Jun 26, 2025
    Had a lot of fun chatting with @armand_ruiz on stage today at @VentureBeat Transform conference. Well worth checking out.
    user avatar
    VentureBeat
    @VentureBeat
    Jun 25, 2025
    Designing for autonomy isn’t a tweak—it’s a transformation. In “Designing for Autonomy: How Agentic AI Will Reshape Enterprise Architecture,” we explored how businesses must reimagine their tech stacks to support agents that can observe, decide, and act without human prompting.
    6.9K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Mar 26, 2023
    If you want to try the "Dolly" LLM model (GPT-J-6B + Alpaca)- here is a Colab with a LoRa finetuning of it. It will need a 3090 or A100 to run.
    Colab logo
    colab.research.google.com
    Dolly - LoRA - GPT-J-6b + Cleaned Alpaca Dataset.ipynb
    Colaboratory notebook
    76K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Mar 4, 2023
    The new Flan-UL2 20B released this week is a big step forward in open sourced models. I put together a Colab rli.to/TA7Y3 walking through loading the @huggingface version in 8bit for people to play with and try the full 2048 token span. Congrats to @YiTayML and co.
    82K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Feb 21, 2024
    Gemma - 2B 7B 6T A baby Gemini and the first open LLM model from Google since UL2-20B Let the Fine Tuning begin!! youtu.be/W7xYXCQQvss Also looking forward to a multi-lingual version since many in Google are passionate about languages beyond English
    25K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Feb 18, 2024
    Gemini 1.5 Pro is exceptionally good at summarization and extraction on long form content like a whole book. In this video I look at making a book summary and getting resources out of it in 270k tokens youtu.be/Pvk4vqescz4 #GoogleAIstudio #BuildWithGemini
    Gemini 1.5 for Summarization
    From youtube.com
    33K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    May 10, 2018
    The wonderful art of @hardmaru and latent spaces on the main stage at #io18
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Feb 19, 2024
    How well does Gemini 1.5 Pro do at Video Analysis? In this vid I get it to go through 800k of video tokens from a recent Ken Kennedy Lecture by @JeffDean See what it really shines at including needles in the videostack youtu.be/pt78XWrOEVk #GoogleAIstudio #BuildWithGemini
    98K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Jul 4, 2024
    Checkout RouteLLM - a great helper for saving money on tokens using matrix factorization for routing models. Cool to see even more innovations coming out of @lmsysorg
    16K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Dec 15, 2023
    New Video Up! Using Gemini Pro with @langchain - LCEL Chains, Mini RAG, PAL + Multimodal youtu.be/G3-YOEVg-xc #BuildWithGemini
    25K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Dec 11, 2024
    Apart from being a very strong model Gemini 2.0 Flash is a huge step forward in multimodality. Bidirectional Streaming Voice & Video unlocks many new opportunities for apps and agents. I have been playing with it and have put together a video (link in reply) #Gemini2
    12K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Feb 25, 2024
    Using Gemini 1.5 Pro for Code Analysis and Generation with the CrewAI codebase youtu.be/I3YE1ESKiN8
    21K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Jan 29, 2025
    I don’t think you get to claim credit for someone else’s paper by saying “if we had published we would have shown we did it first”
    user avatar
    Mark Chen
    @markchen90
    Jan 28, 2025
    Congrats to DeepSeek on producing an o1-level reasoning model! Their research paper demonstrates that they’ve independently found some of the core ideas that we did on our way to o1.
    4.2K
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    Jun 17, 2019
    For people who asked what books I recommended for getting started in Deep Learning here is the slide. @aureliengeron ‘s new book is most up to date with a #Tensorflow2 focus, along with that both books by @fchollet and @iamtrask are great reads 4 anyone startingout #pyconth2019
  • user avatar
    Sam Witteveen
    @Sam_Witteveen
    May 25, 2023
    If you are in Singapore - tonight is Multimodal night at the MLS Meetup. I'm giving a talk on building multimodal agents with Palm2 + @langchain + @huggingface hub - still a few spots free - 7pm @DevSpaceSG meetup.com/machine-learni…
    7.6K