Does memory for AI agents basically mean -

- Save everything to disk. Index it or store in vectorDB. - Search the storage for similarity based on the new prompt - include any finding with the new prompt as system/user prompt (or if you find the exact answer skip the llm call)

?

Or is there more to it ?

Different implementation details have different way to rank or “forget” or supersede facts but that’s pretty much it

Given all the hype that LLMs have got and the valuations these AI companies are getting, I am disappointed at how even the "best" coding agents degrade over time.

WTF - I need to implement a review command to guide to do its job properly.

Can you imagine any other industry charging people money for a product like this ?

When you are charging people money - scratching the surface cannot be an excuse.

There’s a saying for this - “don’t ship your first draft”.

Its first attempt WILL be crap. You can refine its output in the loop as a human, or let it go and then do a quality pass post-hoc.

The beauty of the latter is you can codify and automate a subset of the work, and the agent can loop until those bars are cleared then you get a higher quality work product to look at.

Rather than switch stacks, better way to phrase it is that one should try to increase breadth and depth of knowledge.

By breadth I mean learning other programming language, frameworks, system design etc.

By depth I mean that - if you claim to know React - then get to know the deep detail of React, TS and the related ecosystem.

You can showcase the expertize by writing code, demoing it or sharing on github. Writing about it is another way

Planning to quit by end of the year. That is if I dont get fired earlier.

Place is a shithole. CEO/founders not being honest. Chief architect in India.

Feel that current architecture is broken and will not work.

Privately working on a prototype to demo to CEO/founders. Plan to tell the CEO this is what is possible and changes need to happen. Otherwise it is "sayonara"

Happy to see a post on Java and a book as well.

Will checkout the book.

Recovering the application from failures especially when updating multiple data sources, once and only once execution and such things are in the application domain. They have never been done by relational databases. That is the problem solved by the Python SDK of DBOS ( and typescript SDK)

It's sort of a combination of both. The library solves those problems by storing specific data in the database and then taking advantage of the database's ACID properties and transactions to make the guarantees.

Then the DBOS cloud platform optimizes those interactions between the database and code so that you get a superior experience to running locally.

Good to see alternatives coming up to ridiculously complex serverside programming frameworks like SpringBoot, Node, etc etc.

How is Node ridiculously complex? It’s quite sparse.

Agree that Node is easier than SpringBoot, But Writing a simple app is easy. Not so for transactional app that needs to scale and be reliable and be fast. Add to it: docker Kubernetes Provisioning on cloud Log collection Observability ....... and much more

You need a team of several engineers ?