marktolson

It's an easy way to compare the coding and creative strengths of models. I prefer them over reading a tabular comparison of benchmarks which you have no real insights into.

Looks cool but I just ran the results url through gpt and asked it to create a plan for better optimisation. What would the point of the upsell be?

True, but resetting context to avoid bias doesn't require different agent personas with limited scope, it requires refreshing the context which can be done with a single agent.

I got it to transcribe this: "Create tests and ensure all tests pass" and instead of transcribing exactly what I said it outputs nonsense around "I am a large language model and I cannot create and execute tests".

Other than that issue I like it.

Genuinely impressed with this! Thank you!

Crashes on Mac trying to open up a fairly small workspace (max 97,000 rows).

Looks good but there is no way I'm installing powershell as a dependency.

I represent one of those stars. N8N is an amazing system. I use it daily and have seen incredible progress made consistently over the past few years. Currently my team use it for processing millions of workflows each year on a very small server and it hasn't cost us a cent so I would give them more than 1 star if I could. I wouldn't build a new system without it to be honest.

Not sure if you read the description but I clearly stated that I wanted no middleman, so I'm not sure how this solves my use case. Also not sure how it requires the same amount of time, and 'manual intervention every time'? You provision a remote server, install docker then update the local config file and it's done. One line deployments everytime. And why would I update the code inside the container when you need a build process to install dependencies? If you want to update code while the app is running then restart, then docker is not the right solution.

Additionally there are a million different and better ways to deploy services, this suits the use case I described.