corpus.blog
41554 blogs · [ { "id": "01a0d036-4fab-739c-9607-079d046904c2", "title": null, "url": "https://finbarr.ca", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d166800aa", "title": null, "url": "https://finbarr.ca/advice", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1669c60a", "title": "AI research jobs", "url": "https://finbarr.ca/aijobs", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d16bd24de", "title": "Best practices in training neural networks", "url": "https://finbarr.ca/best-practices", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d17b0d008", "title": null, "url": "https://finbarr.ca/blog/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d17bbd9c5", "title": null, "url": "https://finbarr.ca/books", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1869f6ff", "title": null, "url": "https://finbarr.ca/development", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d18dc0102", "title": null, "url": "https://finbarr.ca/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d19a92946", "title": "LLM Performance Tools", "url": "https://finbarr.ca/llm-performance", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1a8a0bc6", "title": "Ideas I've been wrong about", "url": "https://finbarr.ca/mistakes/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1b0df95c", "title": "Performance Notes", "url": "https://finbarr.ca/notes/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1b3a7213", "title": null, "url": "https://finbarr.ca/portfolio", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1bda2c9c", "title": "Prompt engineering examples", "url": "https://finbarr.ca/prompt-engineering", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1cc751b2", "title": "Requests for startups", "url": "https://finbarr.ca/rfs/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1cf0bb82", "title": "Talks I've given", "url": "https://finbarr.ca/talks", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1dda37b8", "title": null, "url": "https://finbarr.ca/transformers/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1e90a99e", "title": "Work advice", "url": "https://finbarr.ca/work", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1eff858d", "title": null, "url": "https://finbarr.ca/articlepromo/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1fccbb64", "title": "Overview", "url": "https://finbarr.ca/talks/useR-2016/", "published_at": "2026-06-08T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae519dc4957", "title": "Latency numbers everyone should know", "url": "https://finbarr.ca/numbers-you-should-know/", "published_at": "2026-04-15T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51acf44b5", "title": "What, exactly, is the GIL?", "url": "https://finbarr.ca/what-is-the-gil/", "published_at": "2026-04-10T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51b813b36", "title": "Making RL Fast", "url": "https://finbarr.ca/making-rl-fast/", "published_at": "2026-04-03T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51badd5a8", "title": "The Bitter Lesson", "url": "https://finbarr.ca/the-bitter-lesson/", "published_at": "2025-06-26T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51c64adec", "title": "Request for research: Monte Carlo Tree Search for reasoning, with PUCT", "url": "https://finbarr.ca/request-for-research-puct/", "published_at": "2025-01-21T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51ca48af7", "title": "RESP advice", "url": "https://finbarr.ca/resp-advice/", "published_at": "2024-11-08T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51d522d85", "title": "Installing Docker on a new VM", "url": "https://finbarr.ca/install-docker-on-ubuntu/", "published_at": "2024-02-02T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51d80b212", "title": "Five years of GPT progress --- Amii talk", "url": "https://finbarr.ca/five-years-of-gpt-talk/", "published_at": "2023-05-29T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51e1cce08", "title": "Deriving the DALL-E lower bound", "url": "https://finbarr.ca/deriving-the-dall-e-lower-bound/", "published_at": "2023-04-05T00:00:00+00:00" }, { "id": "01a07df2-f125-7180-895f-3ae51e9bdc19", "title": "Five years of GPT progress", "url": "https://finbarr.ca/five-years-of-gpt-progress/", "published_at": "2023-03-27T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d084755ca", "title": "How is LLaMa.cpp possible?", "url": "https://finbarr.ca/how-is-llama-cpp-possible/", "published_at": "2023-03-16T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d087936bc", "title": "A step towards self-improving LLMs", "url": "https://finbarr.ca/self-improving-LLMs/", "published_at": "2023-03-07T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0970b735", "title": "Papers I've read this week (March 4th, 2023)", "url": "https://finbarr.ca/papers-ive-read/", "published_at": "2023-03-04T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0a353ff3", "title": "The Sigmoid: a metaphor for technological progress", "url": "https://finbarr.ca/the-sigmoid/", "published_at": "2023-03-02T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0a7bd45e", "title": "Large language models aren't trained enough.", "url": "https://finbarr.ca/llms-not-trained-enough/", "published_at": "2023-02-27T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0afe4c0b", "title": "A pure Python (well, Numpy) implementation of back-propagation", "url": "https://finbarr.ca/backprop/", "published_at": "2023-01-29T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0b003c33", "title": "Pointer Networks", "url": "https://finbarr.ca/pointer-networks/", "published_at": "2017-09-20T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0b995a84", "title": "Do deep networks generalise or just memorise?", "url": "https://finbarr.ca/do-dnns-generalise-or-memorize/", "published_at": "2017-07-04T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0bdfe6da", "title": "Outrageously Large Neural Networks: The sparsely-gated Mixture-of-Experts layer", "url": "https://finbarr.ca/outrageously-large-networks/", "published_at": "2017-07-01T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0bf352bb", "title": "Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour", "url": "https://finbarr.ca/imagenet-1-hour/", "published_at": "2017-06-20T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0c539f0e", "title": "Random Search for Hyper-Parameter Optimization", "url": "https://finbarr.ca/random-search-hyper-parameter-optimization/", "published_at": "2017-03-01T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0d416573", "title": "Useful Bash One-liners", "url": "https://finbarr.ca/Useful-oneliners/", "published_at": "2017-01-20T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0e2b48ab", "title": "A Deep Hierarchical Approach to Lifelong Learning in Minecraft", "url": "https://finbarr.ca/minecraft/", "published_at": "2017-01-03T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0e89e02b", "title": "Larry Ellison on consulting costs", "url": "https://finbarr.ca/softwar-academic/", "published_at": "2016-12-06T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0f728a36", "title": "AlexNet-level accuracy with 50x fewer parameters and <0.5mb model size", "url": "https://finbarr.ca/SqueezeNet/", "published_at": "2016-11-10T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d0f920f25", "title": "Conditional image synthesis with auxiliary classifier GANs", "url": "https://finbarr.ca/conditional-image-synthesis/", "published_at": "2016-11-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d104dc105", "title": "Generative Adversarial Imitation Learning", "url": "https://finbarr.ca/Generative-adversarial-imitation-learning/", "published_at": "2016-11-08T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d108f4932", "title": "Minimal example of how to do model selection in Python", "url": "https://finbarr.ca/ridge-regression-with-cross-validation/", "published_at": "2016-10-26T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d118190d3", "title": "Representation Learning: A Review and New Perspectives", "url": "https://finbarr.ca/representation-learning-overview/", "published_at": "2016-10-20T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1192d231", "title": "Full Resolution Image Compression with Recurrent Neural Networks", "url": "https://finbarr.ca/image-compression-rnn/", "published_at": "2016-10-19T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d11ea4813", "title": "Generative Adversarial Networks and Actor-Critic methods", "url": "https://finbarr.ca/generative-adverserial-networks-and-actor-critics/", "published_at": "2016-10-19T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d125bc1eb", "title": "Using simulated data to train robots", "url": "https://finbarr.ca/Progressive-networks-training-robots/", "published_at": "2016-10-18T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d1326ebd3", "title": "Safe and Efficient Off-Policy Reinforcement Learning", "url": "https://finbarr.ca/Off-Policy-Reinforcement-Learning/", "published_at": "2016-10-18T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d14138f63", "title": "XGBoost: A scalable tree boosting system", "url": "https://finbarr.ca/xgboost/", "published_at": "2016-09-20T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d14de93d4", "title": "Excellent description of how hashtables work", "url": "https://finbarr.ca/excellent-description-of-hashtables/", "published_at": "2015-08-15T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d14dec600", "title": "Full example for using JSONcpp on Unix", "url": "https://finbarr.ca/jsoncpp-example/", "published_at": "2014-09-06T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d14ec87e7", "title": "ARIMA, ARMA, what's the difference?", "url": "https://finbarr.ca/arima-arma-what/", "published_at": "2014-04-21T00:00:00+00:00" }, { "id": "01a0d036-4fab-739c-9607-079d15a6b5ba", "title": "Solving Partial Autocorrelation Functions", "url": "https://finbarr.ca/solving-pacf/", "published_at": "2014-03-03T00:00:00+00:00" } ] posts Claim your blog
Back to Finbarr Timbers
Blog · corpus.blog/blogs/finbarr.ca/posts

Finbarr Timbers

finbarr.ca

2026

https://finbarr.caoriginal ↗
8 Jun 2026
https://finbarr.ca/adviceoriginal ↗
8 Jun 2026
AI research jobsoriginal ↗
8 Jun 2026
Best practices in training neural networksoriginal ↗
8 Jun 2026
https://finbarr.ca/blog/original ↗
8 Jun 2026
https://finbarr.ca/booksoriginal ↗
8 Jun 2026
https://finbarr.ca/developmentoriginal ↗
8 Jun 2026
https://finbarr.ca/original ↗
8 Jun 2026
LLM Performance Toolsoriginal ↗
8 Jun 2026
Ideas I've been wrong aboutoriginal ↗
8 Jun 2026
Performance Notesoriginal ↗
8 Jun 2026
https://finbarr.ca/portfoliooriginal ↗
8 Jun 2026
Prompt engineering examplesoriginal ↗
8 Jun 2026
Requests for startupsoriginal ↗
8 Jun 2026
Talks I've givenoriginal ↗
8 Jun 2026
https://finbarr.ca/transformers/original ↗
8 Jun 2026
Work adviceoriginal ↗
8 Jun 2026
https://finbarr.ca/articlepromo/original ↗
8 Jun 2026
Overvieworiginal ↗
8 Jun 2026
Latency numbers everyone should knoworiginal ↗
15 Apr 2026
What, exactly, is the GIL?original ↗
10 Apr 2026
Making RL Fastoriginal ↗
3 Apr 2026

2025

The Bitter Lessonoriginal ↗
26 Jun 2025
Request for research: Monte Carlo Tree Search for reasoning, with PUCToriginal ↗
21 Jan 2025

2024

RESP adviceoriginal ↗
8 Nov 2024
Installing Docker on a new VMoriginal ↗
2 Feb 2024

2023

Five years of GPT progress --- Amii talkoriginal ↗
29 May 2023
Deriving the DALL-E lower boundoriginal ↗
5 Apr 2023
Five years of GPT progressoriginal ↗
27 Mar 2023
How is LLaMa.cpp possible?original ↗
16 Mar 2023
A step towards self-improving LLMsoriginal ↗
7 Mar 2023
Papers I've read this week (March 4th, 2023)original ↗
4 Mar 2023
The Sigmoid: a metaphor for technological progressoriginal ↗
2 Mar 2023
Large language models aren't trained enough.original ↗
27 Feb 2023
A pure Python (well, Numpy) implementation of back-propagationoriginal ↗
29 Jan 2023

2017

Pointer Networksoriginal ↗
20 Sept 2017
Do deep networks generalise or just memorise?original ↗
4 Jul 2017
Outrageously Large Neural Networks: The sparsely-gated Mixture-of-Experts layeroriginal ↗
1 Jul 2017
Accurate, Large Minibatch SGD: Training ImageNet in 1 Houroriginal ↗
20 Jun 2017
Random Search for Hyper-Parameter Optimizationoriginal ↗
1 Mar 2017
Useful Bash One-linersoriginal ↗
20 Jan 2017
A Deep Hierarchical Approach to Lifelong Learning in Minecraftoriginal ↗
3 Jan 2017

2016

Larry Ellison on consulting costsoriginal ↗
6 Dec 2016
AlexNet-level accuracy with 50x fewer parameters and <0.5mb model sizeoriginal ↗
10 Nov 2016
Conditional image synthesis with auxiliary classifier GANsoriginal ↗
8 Nov 2016
Generative Adversarial Imitation Learningoriginal ↗
8 Nov 2016
Minimal example of how to do model selection in Pythonoriginal ↗
26 Oct 2016
Representation Learning: A Review and New Perspectivesoriginal ↗
20 Oct 2016
Full Resolution Image Compression with Recurrent Neural Networksoriginal ↗
19 Oct 2016
Generative Adversarial Networks and Actor-Critic methodsoriginal ↗
19 Oct 2016
Using simulated data to train robotsoriginal ↗
18 Oct 2016
Safe and Efficient Off-Policy Reinforcement Learningoriginal ↗
18 Oct 2016
XGBoost: A scalable tree boosting systemoriginal ↗
20 Sept 2016

2015

Excellent description of how hashtables workoriginal ↗
15 Aug 2015

2014

Full example for using JSONcpp on Unixoriginal ↗
6 Sept 2014
ARIMA, ARMA, what's the difference?original ↗
21 Apr 2014
Solving Partial Autocorrelation Functionsoriginal ↗
3 Mar 2014