Shreyansh Lodha

With #VALORANTChampions around the corner I'd like to publicly annouce that I've worked on an unofficial android app for @VLRdotgg play.google.com/store/apps/det… It completely ad free and open source. #VALORANT

this guy has 29 models on huggingface at page 2 ranking. no lab behind him. no sponsorship. $2,000 from his own pocket on GPU rentals. he compressed GLM-4.7 to run on a MacBook and quantized Nemotron Super the week it dropped. all public. all free. nvidia is a trillion dollar company with hundreds of teams but they are not the ones quantizing models middle of the night and pushing them out before sunrise. if nvidia stopped tomorrow their employees stop working. people like @0xSero would not. that is the difference between a paycheck and a mission. @NVIDIAAI you talk about making AI accessible. the people actually doing it are right here. 29 models deep burning their own compute with no ask except more hardware to keep going. you do not need to build another program. just look at who is already building for you. one GPU to this man would produce more public value than a hundred internal sprints. i am not asking for charity. i am asking you to invest in someone who already proved it.

@giulio_leone97 @_Evan_Boyle Oh I see, I guess I was lucky to not experience it! Thanks!

@staticvar_dev @_Evan_Boyle it's full of threads reddit.com/r/GithubCopilo…

The mythical agent month

@giulio_leone97 @_Evan_Boyle What are the changes and where can I read more about them? Couldn’t find any post/news around them

@staticvar_dev @_Evan_Boyle yes , bad... really really bad ... even for pro+

@giulio_leone97 @_Evan_Boyle Wait what did I miss out on? Did they change or update limits?

@_Evan_Boyle agents? we can't run more than 2-3 concurrent requests on copilot cli anymore. it's impossible to work with these new limits. especially as a pro+ user, they make zero sense… do I really need 3 separate subscriptions just to use copilot cli in 3 terminals at the same time?

@avidRaghav It’s not exactly “compression”! Once the gradle is done with its task, build brief extracts the important information. Here is an example screenshot, you can see that all the “noise” that gradle emits is gone!

@staticvar_dev Build-brief supports token compression?

Created this specifically for Android / KMP folks, although this would work for any gradle project. Since RTK doesn’t have first class support for gradle. bb.staticvar.dev

AI is making CEOs delusional

Part 2!

@_Evan_Boyle I think there is a small bug. Currently when I ask it to use any other model to get some tasks done, I am not sure if it counts it as a "premium" request or at least the session summary makes it look like it did not count.

@staticvar_dev Keep the feedback coming! Getting better every day.

I have gotten so much work done in last 2 days with Copilot CLI! It's just so so good, can't wait for A2A communication for fleets to be released! Thanks to @_Evan_Boyle and team for the fantastic work!

Honestly @ArtificialAnlys is not appreciated enough! The amount of information and details they have on artificialanalysis.ai is amazing! Helps you understand how models stack against each other, where a model shines and much more! Thanks a bunch!

I cannot emphasize this enough, these models are really good and the limits in Go plan are very very generous for what you are paying! Companies like @anomalyco actually give me hope! 💖

Github Copilot gives you best bang for the buck and its not even close! I hope they don't nerf it!

@noctus91 @thdxr I think they already did, I checked my OpenCode GO's monthly usage a couple of hours back and the % was reduced.

@thdxr It would be great if you could improve the limits of the Go plan.

we're learning a lot about the inference business and they're learning a lot about us there are optimizations they can make because we can provide some guarantees about the workload reminds me of the toyota stuff where they co-designed products with their vendors

@tomkrcha Weirdly enough, I've been trying to trigger swarm mode again via codex and even VS Code, but it just does not seem to work. Checked it on 1.1.31 and the version before.

@staticvar_dev yes, sir! 🫡

Pencil just crossed 100,000 users! Today we are introducing SWARM mode. A team of AI design agents working with you in parallel. Your autonomous design agency. Download now.

@JorgeCastilloPr MiniMax models have good “personality” but it doesn’t do super well with Kotlin yet, especially when compared to models like Kimi K2.5

Anyone tried MiniMax models for coding mobile apps?

I wish @Kimi_Moonshot 's Moderato plan also gave us access to Kimi Claw! I just can't justify the cost for Allegretto plan for MY usecase. But regardless of that, congratulations to the team for absolutely banger release of K2.5, Loving the model so far!

Hey @thdxr / @opencode, loving the "Go" plan so far, wanted to check if you guys plan to launching a $20 GO plan or $50 BLACK plan. I am running out of current GO plan quicker than anticipated (which is understandable), hence why the question.

No codex no, this was not you. What happened to my boy!!

@tomkrcha We can even trigger SWARM mode via codex?

@staticvar_dev So Codex is supported on canvas as well as via CLI. VSCode Copilot in IDE and OpenCode CLI via MCP as well. Copilot CLI not yet, but will look into it.