Log inSign up
Thomas Wiecki
8,119 posts
Thomas Wiecki profile banner
@twiecki

Thomas Wiecki

@twiecki
Founder of @pymc_labs - The Bayesian Consultancy pymc-labs.com @pymc_devs co-author.
pymc-labs.com
Joined September 2011
1,596
Following
17K
Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • @twiecki
    Thomas Wiecki
    @twiecki
    Sep 1
    Really enjoyed this conversation with @dgerlanc around how agents unlock data science's original premise, by finally making analytics and advanced modeling available to a much larger user-base.
    @dgerlanc
    Dan Gerlanc
    @dgerlanc
    Sep 1
    “A premise we had like, I don't know, 15 years ago, we're like, oh yeah, data science is going to be like this amazing thing that's like going to deliver all this business value. I don't think we have really delivered. I think for a large part we've failed, right?” On the latest
    Image
    00:00
  • @twiecki
    Thomas Wiecki
    @twiecki
    Jul 13
    🔮 pymc_forecast: Bayesian time-series forecasting for @pymc_devs — you write the model, it handles forecasting, backtesting & CRPS scoring. Port of @juanitorduz's numpyro_forecast pip install pymc-forecast 📖 pymc-labs.github.io/pymc_forecast/
  • @twiecki
    Thomas Wiecki
    @twiecki
    Jun 23
    pathmc closes a big gap between Causal and Bayesian methods. Usually the tooling existed for one of the two but now you can define your causal DAG and then take that all the way to a fully Bayesian SCM, modeling the full data generative process (something DoWhy can't do).
    @pymc_labs
    PyMC Labs
    @pymc_labs
    Jun 23
    We're open-sourcing pathmc! Structural causal models + #Bayesian estimation in one package. Define assumptions once, then estimate effects, test identifiability, run interventions, and perform sensitivity analysis. GitHub: dub.sh/2hG8kAu Docs: dub.sh/rvUHI78
    1
  • @twiecki
    Thomas Wiecki
    @twiecki
    Jun 16
    The amount of AI workflow tips in this series is astonishing.
    @hugobowne
    Hugo Bowne-Anderson
    @hugobowne
    Jun 16
    Forget agent skills, forget about subagents, forget OpenClaw, forgot autoresearch, forget ralph loops, forgot about your twitter timeline. @twiecki (@pymc_labs) and I spent 10 hours talking with 16 python, data, ML, and AI builders we trust about what they're actually using,
    Image
  • @twiecki
    Thomas Wiecki
    @twiecki
    Jun 4
    In Claude Code you can just do your own compaction when your context is full but you don't want to lose the session. Just ask Claude to build you a new .jsonl session file based on a previous chat history. Your previous sessions live in ~/.claude.
    3
Advertisement
Advertisement