Oliver Maclaren @omaclaren.bsky.social · 27/07/2026quick testing, but Mistral free tier actually seems like a good option for teaching a module on llm agents? quick api set up, medium 3.5 seems capable enough, from what I can tell rate limits are decent?? 020
Oliver Maclaren @omaclaren.bsky.social · 22/06/2026Has anyone managed to use eg openrouter/free for introducing students to llm agents in teaching? What’s the best free yet passable and generally accessible option for students these days? 130
Oliver Maclaren @omaclaren.bsky.social · 24/05/2026testing qwen3.6:27b in pi studio, now with file navigation/viewer, pane resizing, zen mode, keyboard shortcuts, REPL, simple text/code completion suggestions 040
Oliver Maclaren @omaclaren.bsky.social · 15/05/2026added simple collaborative REPLs to pi studio... 000
Oliver Maclaren @omaclaren.bsky.social · 06/05/2026Linearising consciousness by taking the neural tangent kernel 040
Reposted by Oliver Maclarenkepano @stephango.com · 05/05/2026mature things tend to get pressured into being less weird over time, so if you're starting from scratch your advantage is that you can be very weird if you don't start weird you're probably trying to compete inside the local maximum of something more boring than you could be 115723
Oliver Maclaren @omaclaren.bsky.social · 05/05/2026Finally uploaded a revised version of this preprint on parameter identifiability arxiv.org/abs/2502.04867. Approach called 'invariant image reparameterisation' though 'image' used in mathematical sense not picture sense. Took me a while to revise for various reasons, but fairly happy with it now 1114
Oliver Maclaren @omaclaren.bsky.social · 14/04/2026made a low quality little demo of my pi studio gui thing for pi. see github.com/omaclaren/pi... 100
Oliver Maclaren @omaclaren.bsky.social · 01/04/2026New agent manager: just use a browser with vertical tabs + tmux + ttyd 000
Oliver Maclaren @omaclaren.bsky.social · 30/03/2026I doubt you can take these stats at face value but this a lot more than have cited my academic papers to date… 000
Oliver Maclaren @omaclaren.bsky.social · 27/03/2026inspired by opencode folk showing love to the people of pi, I had a go at slopping up an opencode plug-in version of my pi extension pi-studio...seems to work! see github.com/omaclaren/pi... 100
Oliver Maclaren @omaclaren.bsky.social · 18/03/2026for people of pi (and on mac since uses glimpse). instead of asking models side questions while it does stuff, you answer questions about the code in a side quiz: github.com/omaclaren/pi... not sure if any good yet but another attempt to encourage more active code engagement when using agents 191
Oliver Maclaren @omaclaren.bsky.social · 09/03/2026another little pi extension -- pi-repl (github.com/omaclaren/pi...) 130
Oliver Maclaren @omaclaren.bsky.social · 02/03/2026prettied up pi studio (github.com/omaclaren/pi...) with more syntax highlighting and stuff... no idea if actually useful lol but fun being able to make tools like this so quickly with pi.dev (thanks @mariozechner.at ) 150
Oliver Maclaren @omaclaren.bsky.social · 27/02/2026Me and pi made another pi.dev extension…this time ‘pi studio’: opens a (local) browser workspace for annotating model responses/files, running edited prompts, and requesting critiques www.npmjs.com/package/pi-s... 010
Oliver Maclaren @omaclaren.bsky.social · 23/02/2026pi + codex + zed is the most peaceful workflow for me at the moment. Codex is less sneaky and makes less dumb mistakes than claude, pi lets you add your own little simple workflow tweaks and tools easily, and zed is nice and simple and fast for actually reading code 250
Oliver Maclaren @omaclaren.bsky.social · 18/02/2026Got pi coding agent to make itself an extension to render markdown replies with equations in the terminal or browser www.npmjs.com/package/pi-m...npmjs.com 110
Oliver Maclaren @omaclaren.bsky.social · 28/01/2026I missed gh copilot suggest after they deprecated it so here’s a simple bash function to use claude code etc to suggest executable shell commands like it used to gist.github.com/omaclaren/e1... Same deal as eg Simon Willison’s llm-cmd but uses your clis instead of apisgist.github.comMinimal bash function (add to .bashrc) to generate executable shell commands from natural language. Uses claude code in non-interactive mode (with -p). Can be adapted to other clis easilyMinimal bash function (add to .bashrc) to generate executable shell commands from natural language. Uses claude code in non-interactive mode (with -p). Can be adapted to other clis easily - ccc.sh 000
Oliver Maclaren @omaclaren.bsky.social · 19/01/2026Reading the world beyond your head and it’s talking about my claude code usage 000
Oliver Maclaren @omaclaren.bsky.social · 15/01/2026What's the best 'agentic CLI' tool that is easy to set up with a free tier (no cc details or initial credit req.) for students to use in a university course? Main path I can think of is github copilot student developer pack + e.g. opencode (also just agent mode in vscode...)? 230
Reposted by Oliver MaclarenDarren Wilkinson @darrenjw.bsky.social · 01/01/2026Some lecture notes on spatial and temporal modelling, illustrated using R - darrenjw.github.io/spatio-tempo... - #rstats #rspatial #quartodarrenjw.github.ioSpatial and temporal statistics 55017
Reposted by Oliver MaclarenSimon Willison @simonwillison.net · 23/11/2025Olmo 3 is notable as a "fully open" LLM - all of the training data is published, plus complete details on how the training process was run. I tried out the 32B thinking model and the 7B instruct models, + thoughts on why transparent training data is so important simonwillison.net/2025/Nov/22/...simonwillison.netOlmo 3 is a fully open LLMOlmo is the LLM series from Ai2—the Allen institute for AI. Unlike most open weight models these are notable for including the full training data, training process and checkpoints along … 219134
Oliver Maclaren @omaclaren.bsky.social · 06/09/2025Any really good actually insightful or helpful articles or resources on teaching/learning in the age of LLMs? 110
Oliver Maclaren @omaclaren.bsky.social · 24/06/2025Some misc observations on trying to use LLMs to help on a current research project… 100
Reposted by Oliver Maclarenivosiekmann.bsky.social @ivosiekmann.bsky.social · 29/05/2025Can we infer chemical reaction networks from time series data? Can we observe concentrations of substances over time and infer which substances react with each other? Learn if (and how much) this is possible from Yong See Foo, @adrianazanca.bsky.social and JenniferFlegg arxiv.org/abs/2505.15653arxiv.orgQuantifying structural uncertainty in chemical reaction network inferenceDynamical systems in chemistry and biology are complex, and one often does not have comprehensive knowledge about the interactions involved. Chemical reaction network (CRN) inference aims to identify,... 1133
Oliver Maclaren @omaclaren.bsky.social · 05/05/2025I tried defining a simple ‘prompt script language’ that sort of works and is fun to play with gist.github.com/omaclaren/b5...gist.github.comA simple 'prompt script language' (PSL) specification, giving instructions for an llm to allow prompting in a pseudocode-like wayA simple 'prompt script language' (PSL) specification, giving instructions for an llm to allow prompting in a pseudocode-like way - psl.md 000
Reposted by Oliver MaclarenFred Scharmen @sevensixfive.bsky.social · 26/01/2025I love you bluesky but I admit I don’t totally understand the rabid and uncritical AI hate that predominates here 4631433
Oliver Maclaren @omaclaren.bsky.social · 15/04/2025The voice mode for all the llms keeps trying to sound more human but I kinda just want my robot friends to sound like robots 150
Oliver Maclaren @omaclaren.bsky.social · 04/04/2025I’ve got some students who wanna play around with eg 70B open llms. Can’t run locally obvs. What’re their best options in terms of ease of use and cost? Mainly just running for now but also maybe fine tuning at some point 210
Oliver Maclaren @omaclaren.bsky.social · 01/04/2025No matter how much beef you have with someone you don’t need to do this shit. Crediting people appropriately is allowed 160
Oliver Maclaren @omaclaren.bsky.social · 25/03/2025Here’s a thing. A bit bitter sweet for a few reasons, and also my name is spelled wrong lol but still www.sciencedirect.com/journal/jour... 6293
Reposted by Oliver MaclarenEugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 09/03/2025Smashing the endorse button as fast as I can www.lesswrong.com/posts/oKAFFv...lesswrong.comA Bear Case: My Predictions Regarding AI Progress — LessWrongThis isn't really a "timeline", as such – I don't know the timings – but this is my current, fairly optimistic take on where we're heading. … 108719
Oliver Maclaren @omaclaren.bsky.social · 26/02/2025Considering switching to cursor solely because the grammarly desktop widget thing works in it but not vscode 020
Oliver Maclaren @omaclaren.bsky.social · 25/02/2025Claude is back 🙏 but also even anthropic is still at this point 220
Oliver Maclaren @omaclaren.bsky.social · 24/02/2025This came up recently but I’m not 100% sure how to think about it. Eg suppose you have a nonlinear function of two variables f(x,y) evaluated on a 2D grid. One apparently useful way to approximate it is to think of the values as entries in a matrix and use matrix factorisation. But why? 100
Oliver Maclaren @omaclaren.bsky.social · 18/02/2025Idle stats thought — calling them confidence subsets/subintervals instead of confidence sets/intervals would more honestly indicate that the coverage guarantees hold with respect to some set of initial of values considered 050
Reposted by Oliver MaclarenMelanie Mitchell @melaniemitchell.bsky.social · 13/02/2025Do large language models develop "emergent" models of the world? My latest Substack posts explore this claim and more generally the nature of "world models": LLMs and World Models, Part 1: aiguide.substack.com/p/llms-and-w... LLMs and World Models, Part 2: aiguide.substack.com/p/llms-and-w...aiguide.substack.comLLMs and World Models, Part 1How do Large Language Models Make Sense of Their “Worlds”? 1421158
Oliver Maclaren @omaclaren.bsky.social · 10/02/2025Hello, here is some new work called ‘Invariant Image Reparameterisation: A Unified Approach to Structural and Practical Identifiability and Model Reduction’. See arxiv.org/abs/2502.04867 261
Oliver Maclaren @omaclaren.bsky.social · 07/02/2025Shout out to this random book exchange setup I found while taking my daughter for a walk! 1121
Oliver Maclaren @omaclaren.bsky.social · 23/01/2025A nice summary on the other site of RL self improvement by @ptrschmdtnlsn.bsky.social “we keep distilling what we conclude after a lot of thinking into what we conclude intuitively in a single step of thinking, which in turn improves what we conclude with a lot of thinking, and so on” 130
Oliver Maclaren @omaclaren.bsky.social · 22/01/2025Just finished @melaniemitchell.bsky.social ‘s really nice book on AI. Just pre- chatgpt etc but good coverage of context leading up to recent explosion. Based on this quote I’d say at least 25 years to ‘AGI’. Will be interesting to see… 131
Oliver Maclaren @omaclaren.bsky.social · 21/01/2025Man still only o1 can solve this. Not Claude, not DeepSeek not Gemini … (DeepSeek app can’t even extract the words from the image correctly…had to pass manually) 010
Reposted by Oliver MaclarenMichael Plank @michaelplanknz.bsky.social · 16/01/2025We are advertising two PhD scholarships to work on models of the interaction between ethnicity and infectious disease dynamics at the University of Canterbury in New Zealand. Open to international students. Please share! www.findaphd.com/phds/project... @iddjobs.bsky.socialfindaphd.comModelling the interaction between ethnicity and infectious disease dynamics at University of Canterbury on FindAPhD.comPhD Project - Modelling the interaction between ethnicity and infectious disease dynamics at University of Canterbury, listed on FindAPhD.com 02524
Reposted by Oliver MaclarenPedro Sant’Anna @pedrosantanna.bsky.social · 12/01/2025I have been trying to get more familiar with the basics and advances Reinforcement Learning. I wonder if anyone have seen a paperthat use potential outcome notation within the RL world. This would make the life of many economists much easier. Anyone? All help is appreciated! 95112
Reposted by Oliver MaclarenIlya Chugunov @ilyac.info · 10/01/2025My ACM Crossroads article is out now! Should be a fun read if you're interested in a broad overview of my computational photography work: dl.acm.org/doi/10.1145/...dl.acm.orgThe Inverse Problems You Carry in Your Pocket | XRDS: Crossroads, The ACM Magazine for StudentsIn the spaces between data-hungry generative models and measurement-rich computational imaging, we can find the field of computational photography. Can cell phone cameras be an accessible and affordable bridge between modern computer vision and ... 062