This is similar to the tool call (fixed code & dynamic params) vs code generation (dynamic code & dynamic params) discussion: tools offer contrains and save tokens, code gives you flexibility. Some papers suggest that generating code is often superior and this will likely become even more true as language models improve
I am working on a model that goes a step beyond and even makes the distinction between thinking and code execution unnecessary (it is all computation in the end), unfortunately no link to share yet
What you're missing is how to use the tools properly. With solid documentation, good project management practices, a well-organized code structure and tests, any junior engineer should be able to read up on your codebase, write linted code following your codebase style, verify it via tests and write you a report of what was done, challenges faced etc. State of the art coding agents will do that at superhuman speeds.
If you haven't set things up properly (important info lives only in people’s heads / meetings, tasks dont have clear acceptance criteria, ...) then you aren't ready for Junior Developers yet. You need to wait until your Coding Agents are at Senior level.
I want to use browser-use in Cursor but I am using another option because it doesn't support MCP integration which is the common language they support for external tools
The closest alternatives in this space would be allennlp [1], the recently released pytext [2] and spacy [3]. pytext's authors wrote some comparison on the accompanying paper [4] and this GitHub issue [5].
We have done extensive testing in the context of chatbot intent classification and in our particular problem nothing (including CNN, LSTM, fasttext plus LUIS, Watson and other proprietary classifiers) has been able to beat a simple linear model trained on char n-gram features.
fast.ai is great and I also want to love it but I find the coding style extremely unconventional and uninviting to the point that I only use the library when there is really no better alternative (ex for the lr finding features). Variable naming and wildcard imports would be my main complains. But thanks a lot Jeremy for the course and for posting the style guide.
We evaluated using ConceptNet Numberbatch but in the end went with fasttext because of the treatment of OOV words using sub word information. This is important for us because we work with Social Media where misspellings are very frequent and we have found this helps a lot. Are you also looking into these sort of enhancements? How do you usually deal with OOV words?
If you follow the right people in your fields of interest you will have an endless stream of interesting updated content. You can't personalize HN to have the exact sort of news you want but you can tune the people you follow in Twitter and get the best source of curated news there is for you. I frequently find stuff like new library releases/interesting blog posts from someone I follow well before they appear (if they ever do) in places like HN/Reddit subs I frequent.
Great work! A question: how do we scale the computations? I know how long my experiments normally take on a 40 core PC. How long will it take and how much will it cost if I use Empirical?
A study assessing 6 months (n=112) and 12 months (n=96) of melatonin treatment of 2mg in a controlled release capsule, taken 1-2 hours prior to sleep, people aged 20-80 with primary insomnia failed to show any tolerance to the treatment. The authors noted a slight sensitization to the effects of melatonin at the 3-4 month period, which was attributed to better entrainment of the circadian rhythm. These results have been replicated in another study, lasting 6 months, with a sample of 791 people. No melatonin tolerance due to usage was observed. Another study, lasting 6 months with a sample size of 421 people also replicated these results.
https://github.com/elyase/awesome-personal-ai-assistants?tab...