Such a pleasure to host my friends Bill Engels and Jesse Grabowski, to talk #GaussianProcesses and #AIworkflows no less! Enjoy, my dear Bayesians 🖖

Pierre-Simon Laplace@learnbayesstats.bsky.social · last wk.

New episode of LBS is out! @alex-andorra.bsky.social is joined by Bill Engels & Jesse Grabowski to discuss Gaussian processes, kernels, PyTensor, and PTGP, including HSGP, computational efficiency, AI-generated statistical code & more! 🔗Watch the full episode: learnbayesstats.com/episode/gaus...

We kick of the visiting speakers' research seminar today at @princeton.edu with @alex-andorra.bsky.social joining us for a talk and workshop on #BayesianWorkflow and AI-assisted statistical modeling. These sessions are open to the public on zoom. For details, see the flyer below!

Huge pleasure to host my friend Alexander Fengler on the show, to go through his latest work on #AmortizedInference and the power of the neural nets + Bayes combination!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 3w ago

What if we learned the likelihood instead of the posterior? In Episode 165, Alex Fengler joins @alex-andorra.bsky.social to discuss HSSM, likelihood approximation networks, amortized inference and simulation-based inference and more ... 🔗 learnbayesstats.com/episode/hssm... #Bayesian #PyMC

Today's episode is a #BayesianBlockbuster, where I talk with three of the most influential teachers of the #stats world... plus a surprise somewhere in the middle 😏 It was an absolute blast to record it -- I hope it'll be as good to listen to it for you!

Pierre-Simon Laplace@learnbayesstats.bsky.social · last mo.

In this special episode of Learning Bayesian Statistics, @alex-andorra.bsky.social is joined by Andrew Gelman, Aki Vehtari, and Richard McElreath to discuss Bayesian Workflow. From simulation and hierarchical pooling to causal inference, there’s a lot to unpack 🎧 lnkd.in/geX2QkxV #Bayesian

New episode is out! A very practical one, where we dive into the new research by Adrian Seyboldt, Eliot Carlson and Bob Carpenter about how to make your MCMC samplers go faster -- lots of gems in there, make sure to tune in!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 2mo ago

How can we make Hamiltonian Monte Carlo sample faster? In this episode, @alex-andorra.bsky.social , Adrian Seyboldt, and Eliot Carlson dive into HMC preconditioning, mass matrix adaptation, and Nutpie - including a median 4× speed-up across 114 models 🎧 learnbayesstats.com/episode/fast...

Some really interesting insights about #Bayes & #LLMs in this new episode -- hope you'll enjoy!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 2mo ago

Bayesian methods aren't going away in the age of LLMs. Christopher Krapu joins @alex-andorra.bsky.social to discuss GPUs, Gaussian Processes, probabilistic AI and more! 🎧 learnbayesstats.com/episode/baye... #bayesian #GPU #AI #LLM #Gaussianprocess #probablisticai

Episode 161 of the show is out, with @lacerbi.bsky.social ! We dive into all things neural processes, fast #Bayesian inferences, and LLMs -- hope you enjoy 🖖

Pierre-Simon Laplace@learnbayesstats.bsky.social · 3mo ago

Episode 161 is out 🎧 In which @lacerbi.bsky.social explains why transformers are secretly neural processes, how his Amortized Conditioning Engine unifies inference and prediction, and why "amortize everything" needed a rethink. 🔗 learnbayesstats.com/episode/161-... #bayesian #bayesianinference

Such a blast to come on the show @vadenmasrani.bsky.social @incrementspod.bsky.social ! Always fun to nerd out on #Bayesian topics to start a day :)

Increments Podcast@incrementspod.bsky.social · 3mo ago

New ep! The haters of Bayesian epistemology @vadenmasrani.bsky.social and Benny spend some time defending Bayesian statistics alongside the Stats God @alex-andorra.bsky.social. The Reverend is finally getting his comeuppance on Increments www.youtube.com/watch?v=7BZV...

🎙️ New episode! @alex-andorra.bsky.social sits down with Matthijs Hollanders on Bayesian occupancy models for wildlife data - what they are, why camera traps break classical approaches, and how his occARU R package handles it with hierarchical GPs and shrinkage priors. 🔗 lnkd.in/dw3WuMBg #bayes

Check out the new baygent skill @alex-andorra.bsky.social and I wrote to automate your amortized workflows! 🚀 learnbayesstats.com/blog-posts/a...

Train Once, Infer a Million Times: Bayesian Inference That's (Finally) Cheap at Query Time

MCMC re-runs from scratch on every new dataset. Amortized Bayesian inference pays the cost once -- then gives you a posterior in milliseconds, forever. Stefan Radev (BayesFlow creator) and I built an ...

learnbayesstats.com

Getting back to the basics with this one! What even *is* #BayesianStats ?? Prepare to get your mind blown by Vaden Masrani...

Pierre-Simon Laplace@learnbayesstats.bsky.social · 3mo ago

🎙️ Bayesian epistemology is Bayesian statistics minus the statistics. In ep 160 Vaden Masrani joins @alex-andorra.bsky.social to talk about why Bayes' theorem is great with real data, why it breaks down on one-off future events with nothing to count👇 🔗 lnkd.in/d3v42BU2 #bayesianstatistics

Big update to my #BayesianWorkflow and #CausalInference #AgentSkills ! Now teach PyMC 6 + ArviZ 1.0 first-class -- the latest best practices, baked in -- while still running on PyMC 5 if you can't upgrade yet. Your AI agent should know the newest idioms! github.com/Learning-Bay...

GitHub - Learning-Bayesian-Statistics/baygent-skills: A set of skills to call your agent Bayes. Thomas Bayes.

A set of skills to call your agent Bayes. Thomas Bayes. - Learning-Bayesian-Statistics/baygent-skills

github.com

Second part of my conversation with Stefan Radev is out, my dear Bayesians! We dive deep into how simulations improve #Bayesian workflows, how to do cheap #sensitivity and #multiverse analysis, and where #BayesFlow is headed next. Enjoy!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 5mo ago

Episode 158 is out 🎙️ @alex-andorra.bsky.social sits down with Stefan Radev to talk amortized Bayesian inference, multiverse analysis, and what a foundation model for Bayesian inference should actually look like and more.. 🔗 lnkd.in/dau9_eA7 #BayesianStatistics #AmortizedInference #MachineLearning

Latest episode is out, my dear #Bayesians! A deep dive into #AmortizedInference, what it looks like in practice, and how to teach it to your AI agents. Tune if you wanna see how to do fast, amortized inference that scales -- live, demoed by Stefan 😉

Pierre-Simon Laplace@learnbayesstats.bsky.social · 5mo ago

🎙️ New episode alert! In this episode @alex-andorra.bsky.social & Stefan Radev dive into amortized inference, train a neural net once on sims, deploy on real data as many times as you want. They cover sim-to-real, psych & neuro as test beds, honest failure modes and more ... lnkd.in/dCY85k4g

New episode is out forks! BALD, BED and other fun academic acronyms 😏

Pierre-Simon Laplace@learnbayesstats.bsky.social · 6mo ago

Most stats thinking starts with a dataset. Bayesian experimental design asks: which data should you collect first? Ep 156 of Learning Bayesian Statistics with @alex-andorra.bsky.social and Adam Foster covers: 👉Expected information gain 👉BALD 👉Deep adaptive design and more ... 🎧 lnkd.in/ebjV9xXS

New episode is out! I sit down with Andreas Munk to explore one of the most important -- and most underappreciated -- tensions in modern Bayesian practice: the gap between what we can do in research and what actually gets deployed in the wild. Hope you like it!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 6mo ago

🎙️ New episode of Learning Bayesian Statistics! EP 155 with @alex-andorra.bsky.social & Andreas Munk, why Bayesian inference still hasn't broken into everyday use. The barrier isn't the math, it's the mental shift 🔗 lnkd.in/gchb6bqj #Bayesian #ProbablisticProgramming

⚽ Last week, I was thrilled and honored to present our #SoccerFactorModel to Field of Play 2026 in Manchester! 🎙️ It was an absolute blast meeting all these brilliant people, and I can't thank enough the FoP team, especially Dominic Jordan and John Carney for their trust and invitation! 🧵 Thread 👇

Wrote up the full story behind the #CausalInference agent skill. Key insight: agents *know* causal inference, but they never stop to ask "is my DAG right?" or "does this survive a placebo test?" The skill doesn't add knowledge. It adds discipline. Full breakdown: learnbayesstats.com/blog-posts/c...

I Built an Agent That Refuses to Say "Causes"

Coding agents confidently say "X causes Y" without drawing a DAG, checking assumptions, or running refutation tests. I built an Agent Skill that won't let them. 100% eval pass rate vs 68% without -- t...

learnbayesstats.com

Just released a new #AgentSkill: causal inference that refuses to say "causes" until it's earned the right. DAG before data. Assumptions before estimates. Refutation before claims. 100% eval pass rate vs 68% without the skill. The gap is in the reasoning, not the code. 👉 github.com/Learning-Bay...

baygent-skills/causal-inference at main · Learning-Bayesian-Statistics/baygent-skills

A set of skills to call your agent Bayes. Thomas Bayes. - Learning-Bayesian-Statistics/baygent-skills

github.com

New episode is out, my dear Bayesians! All about #CausalInference, #Experimentation at scale, and #GaussianProcesses -- definitely a fun one!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 7mo ago

New Episode Alert! 🎙️ Scaling #BayesianCausalInference with Thomas Pinder, Netflix & creator of GPJax Essential listening for anyone working at the frontier of Bayes, Experimentation & Causal Inference 📈 🔗 learnbayesstats.com/episode/154-... #Bayesian #JAX #MachineLearning #CausalInference #GPJax

The show now has a blog section 🍾 Check out @alex-andorra.bsky.social 's first post!

Alexandre Andorra@alex-andorra.bsky.social · 7mo ago

I taught my coding agent to think like a Bayesian, and wrote up what I learned. LLMs know #BayesianStats but skip half the workflow. So I built an open-source Agent Skill that enforces the full workflow every time Works with Claude Code, Cursor, Gemini, etc. learnbayesstats.com/blog-posts/b...

2 weeks ago, I published my #BayesianWorkflowSkill, which scored 100% on evals. Then, the brilliant Max Göbel actually used it on a real model. And it broke 🤨 Lesson: evals are necessary, but not sufficient. Get your tools into real hands early! Full blog post: learnbayesstats.com/blog-posts/b...

A Researcher Stress-Tested My Bayesian Agent Skill. Here's What Broke

A researcher stress-tested my Bayesian Agent Skill on a real hierarchical model. It caught the workflow violations but missed a subtle prediction bug, had no sparsity priors, and lacked classification...

learnbayesstats.com

I taught my coding agent to think like a Bayesian, and wrote up what I learned. LLMs know #BayesianStats but skip half the workflow. So I built an open-source Agent Skill that enforces the full workflow every time Works with Claude Code, Cursor, Gemini, etc. learnbayesstats.com/blog-posts/b...

I Taught My Coding Agent to Think Like a Bayesian

Teach your AI coding agent to think like a Bayesian. Discover the bayesian-workflow Agent Skill that prevents subtle statistical errors in PyMC models.

learnbayesstats.com

New episode is out 🍾

Pierre-Simon Laplace@learnbayesstats.bsky.social · 7mo ago

Fundraising = Belief Updating? 🧠📉 New episode is out! I'm talking with Cherian Koshy about the Neuroscience of Philanthropy. We discuss: ✅ Why generosity is hardwired ✅ Solving the Generosity Gap ✅ Cognitive friction ✅ Ethical AI Check it out: 🔗 learnbayesstats.com/episode/neur...

at my job I’m expectedutilitymaxxing. I chatted with @alex-andorra.bsky.social on how to make it scalable and fast in an industrial context on top of the probabilistic programming language you already love, PyMC.

Alexandre Andorra@alex-andorra.bsky.social · 8mo ago

New episode is out, where we dive into how to use #ProbabilisticModel to do #DecisionMakingOptimization -- because the best model is useless if you can't make decision from it. See you in there!

Just published my first open-source #AgentSkill! It's called bayesian-workflow, and helps you do #BayesianAnaylsis the right way -- well, at least I hope... Check it out here: github.com/Learning-Bay...

GitHub - Learning-Bayesian-Statistics/baygent-skills: A set of skills to call your agent Bayes. Thomas Bayes.

A set of skills to call your agent Bayes. Thomas Bayes. - Learning-Bayesian-Statistics/baygent-skills

github.com

Pierre-Simon Laplace@learnbayesstats.bsky.social · 7mo ago

Hellooooooo my dear Bayesians! We just open-sourced an #AgentSkill that teaches coding agents to do #Bayesian stats properly. No more skipped diagnostics, no more point estimates without uncertainty, no more "trace plots look fine". Works with Claude Code, Cursor, Kimi, Gemini CLI, and more

New episode is out, where we dive into how to use #ProbabilisticModel to do #DecisionMakingOptimization -- because the best model is useless if you can't make decision from it. See you in there!

Pierre-Simon Laplace@learnbayesstats.bsky.social · 8mo ago

Episode 152 is out 🎙️ Host @alex-andorra.bsky.social talks with Daniel Saunders about a Bayesian decision theory workflow. Big idea: stop optimizing for model accuracy and start optimizing for decision value. 🔗 lnkd.in/gw_uGaZc #Bayesian #DecisionTheory #DataScience #Optimization

Diffusion models & flow matching are reshaping simulation-based inference. Thus, we wrote the first tutorial review on diffusion-based SBI. For an overview or a deep dive, check it out and let us know what you think: arXiv: arxiv.org/abs/2512.20685 Web: bayesflow-org.github.io/diffusion-ex...

Diffusion Models in Simulation-Based Inference: A Tutorial Review

Diffusion models have recently emerged as powerful learners for simulation-based inference (SBI), enabling fast and accurate estimation of latent parameters from simulated and real data. Their score-b...

arxiv.org

My #AdvancedRegressionModeling course, written with the brilliant Ravin Kumar and @tomicapretto.bsky.social, is now available through my Topmate profile! So do give it a try and let me know what you think in the comments 👇 See you soon in the Intuitive Bayes' Discourse 🖖 topmate.io/alex_andorra...

I don't code anymore—I build. AI lets me ship projects by describing what I want, not how to do it. In 10 days, I shipped 6 projects with OpenCode + Claude Opus 4.5. The gap between idea and product is tiny now. Would you try this? Repost or comment! buff.ly/blroW24 #ai #buildfast

You Can Just Make Stuff with OpenCode and Claude Opus 4.5

written by Eric J. Ma on 2025-12-28 | tags: ai opencode claude automation workflow llm reasoning development review tools

buff.ly