What remains after you fit a model; the unexplained, the interesting leftovers. I am Rodrigo Nemmen. This is where I write about what interests me. For my...
I just released a major new version of BCES , the Python package for linear regression with measurement errors. In version 2.0, the big addition is weighted least squares (WLS) regression. What's new BCES has handled the case where both X and Y have measurement errors since version 1.0, using the bivariate correlated errors and intrinsic scatter method from Akritas & Bershady (1996) . But there's…
Abstract The leading AI models keep getting more expensive, but are they worth it? I benchmarked dozens of LLMs across coding and general reasoning tasks, then mapped their performance against API pricing. The results are striking: a handful of models—most notably DeepSeek V3.2 and Minimax 2.5—deliver 85–95% of frontier performance at a fraction of the cost. Meanwhile, premium models like Claude…
I was tired of manually figuring out which of my pip-installed packages need to be updated while excluding the conda ones, so I just created a tool that takes care of this. You fire pip-select , and it will check which of your pip packages can be updated while excluding the ones managed by conda, with a nice progress bar: Conda environment detected at: /home/user/miniconda3/envs/myenv Detected 84…
With the explosion of Claude Code , developers are struggling with handle the credit limits. I have seen plenty of Reddit posts complaining about the limitations imposed by the Claude Max plan, even though it costs a whopping $200! I access Anthropic’s models using the Poe API service , which also gives me access to models by the other major players in the field. In my experience with CC so far, I…
New year, new science. Time to think about my favorite papers of 2025. This is of course a hugely biased list, convolved with my personal scientific preferences. This list will be focused on the topics of black hole astrophysics. Phenomenology: AGNs, little red dots Not a terribly exciting year on the observational front for AGNs and black hole binaries. If you discount little red dots, that is.…