garybake

"Which AI model should we build around?"

I'd ask a different question: What will it cost us when we're wrong?

Models change. Providers change. Pricing changes.

Good AI architecture doesn't predict the winner. It makes changing your mind cheap.

As LLM providers ship breaking changes on a near-monthly cadence, the difference between a 10-minute .env swap and a 400-line PR comes down to architecture. This post introduces seam-driven design for Python LLM apps - five narrow interfaces (provider, prompt, tools, config, observability) that can each be replaced independently - illustrated with a working FastAPI + LangChain reference app. Readers get a concrete checklist and two hands-on drills they can run against the open-source code immediately.

Same here. I'm using langchain which creates a varchar id column. It also has different collections on the same table.

Magnetic USB cables are a godsend for young kids with tablets.

I've had that on my 'to buy' list for too long. In such a rapidly changing field like computing, books have to be really special to reach the 'classic' accolade.

I'd say its about 1/3 biography, 1/3 history of the early cold war and 1/3 game theory. Neuman definitely deserves a good biography though.

This here. I think the author is confusing what a 10x developer is.

I took the 10x dev to mean somebody who may work efficiently but also improves all of the devs around them. Improving yourself by 2x is one thing but improving the team around you by 2x is where you get the 10x.

"Tape it to the fridge" Please don't do this, at least keep it hidden from general viewing.

Like in moneyball. It could have a bearing on players wages, sponsorship etc. I hope too hope they loose though.

paperswithcode has a link to the repo

https://paperswithcode.com/paper/slide-in-defense-of-smart-a...

I got stuck on this problem a while ago. I was doing reinforcement learning and the environment needed to run in pypy to make it fast enough. I ended up using Pyro4 to bridge pypy (environment) and CPython (keras).

https://www.manning.com/books/deep-learning-and-the-game-of-... This is an amazing book that goes over how alpha go works and how to replicate it. The book uses keras.

You can use it for "what if" analysis. - Take a fork of the database - Transform data in the fork. i.e. costs+5% - Point your current queries to run on the fork

You can see how with just the effort repointing queries a reporting app could show the real world and the world you are modelling.

I always figured 'model zoo' was a general term for for a collection of trained models.

I would recommend micropython on the esp8266 as then you can use the same code that was running on the pi with a few tweeks. https://github.com/micropython/micropython/tree/master/esp82... http://docs.micropython.org/en/latest/esp8266/