Rithesh Kumar

insomnia developer

[Rithesh Kumar @rk625dev]

@encoredotdev Building agentic knowledge base for speeding up the learning curve

[Rithesh Kumar @rk625dev]

@JackWoth98 Provide mcp support to add mcp server directly thru agy cli from terminal commands

[Rithesh Kumar @rk625dev]

@JackWoth98 @rodydavis I want to change and the bring back the plan mode auto accept default all can be chosen by toggling with alt + tab and can bring the old style of the plan preview

[Rithesh Kumar @rk625dev]

@_anshulr Can u bring browser preview in antigravity and ability to point and edit features in frontend can u do that like how cursor has @_anshulr

[Hitesh Choudhary @Hiteshdotcom]

Very soon, CTC will include token cost.

[Rithesh Kumar @rk625dev]

@vipulgupta2048 This is what I wanted great one btw @vipulgupta2048

[Rithesh Kumar @rk625dev]

@teortaxesTex This a great for deepseek users if we get v4.1 in month

[Rithesh Kumar @rk625dev]

@dropstoneio hey the weekly credits got exhausted for just 2 prompts itself no heavy workload

[Rithesh Kumar @rk625dev]

@ivanfioravanti Take a look at @CommandCodeAI built by @MrAhmadAwais it a cli built entirely to solve the tool call issue in deepseek read this threads

[Rithesh Kumar @rk625dev]

@ivanfioravanti Check @CommandCodeAI which built for deepseek

[Rithesh Kumar @rk625dev]

@TimJayas Try @CommandCodeAI

[Rithesh Kumar @rk625dev]

@TimJayas Which plan are u in

[Rithesh Kumar @rk625dev]

permanent pricing reduction for the minimax m3 api here @MiniMax_AI ⚡️⚡️⚡️⚡️

[Rithesh Kumar @rk625dev]

@JackWoth98 @mitgeniusz @antigravity @ntaylormullen These are workarounds why can’t it be native and seamless switch between auto accept edits plan mode default mode and yolo mode it was cool and easy to switch during middle of session like Gemini cli was like @JackWoth98

[Rithesh Kumar @rk625dev]

@_anshulr can u please bring feature of cursor in antigravity. it also there in the ai studio too it would nice if antigravity supports this stuff https://t.co/MpuZJSnnDP

[Cursor @cursor_ai]

With Design Mode, you can now point, draw, or talk to update your UI.

[Rithesh Kumar @rk625dev]

@shengzheyao Can please change the plan mode the way it is implemented in the cli can u bring back the original implementation of Gemini cli how it handled plan , auto accept , normal mode with shift+tab @shengzheyao

[Deli Chen @victor207755822]

🚀 We’re hiring! DeepSeek is forming a new Harness team to build Code Harness from the ground up—may be you can call it DeepSeek Code or something like this hhh🤣🤣🤣

📍 Based in Beijing. Two roles open:
🧠 Harness Product Manager → https://t.co/oliNbSD8xk
👨‍💻 Harness R&D Engineer → https://t.co/xCSK5wSEAR
Research meets product—let's build it together. Hit the links and apply directly! 🔥
#DeepSeek #CodeHarness #AI #Hiring #ProductManager #Engineering #Beijing #Referral

[Rithesh Kumar @rk625dev]

@teortaxesTex Which application are u used for deepseek

[Soso fun ✨ @Soso_fun_yt]

Hello everybody ! ^^

I need to raise a critical issue regarding the @GeminiApp. There is a major compliance and transparency discrepancy regarding the advertised "1 Million Token Context Window" for AI Pro and AI Ultra plans, specifically how it is marketed versus how it actually behaves under the hood during an active chat session. 😅

Google's official documentation and marketing materials clearly state that users get a 1M token context window, boasting the ability to process up to 1,500 pages of text or 30,000 lines of code simultaneously to suggest edits and debug errors. The help center explicitly claims that this larger context window allows Gemini Apps to work with "longer prompts and chats."

However, as a developer, the reality in the actual UI tells a completely different story. While the backend can successfully ingest a massive static file initially on the first prompt, the active conversational memory (the dynamic context window / KV cache for the chat) appears to be severely bottlenecked, dropping significantly to a 16k~ limit. (Or 25-30 messages in average)

As a result, the model quickly suffers from amnesia within the exact same chat session, completely forgetting earlier instructions, code blocks, or constraints.

From a technical and legal standpoint, marketing a 1M token window for "chats" when the interface relies on a highly restrictive active sliding window is misleading.

You cannot advertise a massive dynamic conversational memory based solely on static file ingestion limits.

@joshwoodward, @vikaskansalHQ, and the @GeminiApp Team, we need clear transparency. 🙏

Please update the documentation to clearly differentiate between "Static File Ingestion Context" and "Dynamic Active Conversational Context."

Advanced users and developers need to know the actual active memory limit during a live chat session to avoid massive regressions in their workflows !

From a researcher perspective, fixing this terminology is urgent. ❤️

[Rithesh Kumar @rk625dev]

@RaghavKoch19380 @LexnLin can tell me the process for refund here pls