20 changes found
20 changes found from /v1/search
X post: User: @pauliusztin_ Post ID: 2097282009605218640 Text: The 4 harness components that keep your context window high-signal. https://t.co/twgO51to5d Metrics: {like_count=43,
X post: User: @pauliusztin_ Post ID: 2097282009605218640 Text: The 4 harness components that keep your context window high-signal. https://t.co/twgO51to5d Metrics: {like_count=43,
X post: User: @pauliusztin_ Post ID: 2097282009605218640 Text: The 4 harness components that keep your context window high-signal. https://t.co/twgO51to5d Metrics: {like_count=43,
X post: User: @MetisL2 Post ID: 2094595377018380619 Text: Most AI workflows still rely on a single agent to research, write, and review everything in one context window. ClawUp tak
X post: User: @MetisL2 Post ID: 2094595377018380619 Text: Most AI workflows still rely on a single agent to research, write, and review everything in one context window. ClawUp tak
X post: User: @MetisL2 Post ID: 2094595377018380619 Text: Most AI workflows still rely on a single agent to research, write, and review everything in one context window. ClawUp tak
X post: User: @gippp69 Post ID: 2090140389244326104 Text: KIMI ENGINEERS JUST SHOWED WHY LONG-TERM AGENT MEMORY IS A SYSTEM PROBLEM, NOT A CONTEXT WINDOW PROBLEM the model is only
X post: User: @gippp69 Post ID: 2090140389244326104 Text: KIMI ENGINEERS JUST SHOWED WHY LONG-TERM AGENT MEMORY IS A SYSTEM PROBLEM, NOT A CONTEXT WINDOW PROBLEM the model is only
X post: User: @gippp69 Post ID: 2090140389244326104 Text: KIMI ENGINEERS JUST SHOWED WHY LONG-TERM AGENT MEMORY IS A SYSTEM PROBLEM, NOT A CONTEXT WINDOW PROBLEM the model is only
X post: User: @WesRoth Post ID: 2089291151224557741 Text: Codex can officially use a 1 MILLION token context window with GPT-5.6 Sol now. Set: model_context_window = 1000000 mod
X post: User: @WesRoth Post ID: 2089291151224557741 Text: Codex can officially use a 1 MILLION token context window with GPT-5.6 Sol now. Set: model_context_window = 1000000 mod
X post: User: @WesRoth Post ID: 2089291151224557741 Text: Codex can officially use a 1 MILLION token context window with GPT-5.6 Sol now. Set: model_context_window = 1000000 mod
X post: User: @e_opore Post ID: 2083556853938557308 Text: 20 concepts to master Building AI Agents from Scratch: 𝟏. 𝐋𝐋𝐌 𝐅𝐮𝐧𝐝𝐚𝐦𝐞𝐧𝐭𝐚𝐥𝐬: Tokens, Context Windows 𝟐. ?
X post: User: @e_opore Post ID: 2083556853938557308 Text: 20 concepts to master Building AI Agents from Scratch: 𝟏. 𝐋𝐋𝐌 𝐅𝐮𝐧𝐝𝐚𝐦𝐞𝐧𝐭𝐚𝐥𝐬: Tokens, Context Windows 𝟐. ?
X post: User: @e_opore Post ID: 2083556853938557308 Text: 20 concepts to master Building AI Agents from Scratch: 𝟏. 𝐋𝐋𝐌 𝐅𝐮𝐧𝐝𝐚𝐦𝐞𝐧𝐭𝐚𝐥𝐬: Tokens, Context Windows 𝟐. ?
X post: User: @RituWithAI Post ID: 2078718154289848621 Text: 🚨 The Chinese AI lab that pioneered million-token context windows just launched their own Claude Code. Free. Open sou
X post: User: @RituWithAI Post ID: 2078718154289848621 Text: 🚨 The Chinese AI lab that pioneered million-token context windows just launched their own Claude Code. Free. Open sou
X post: User: @RituWithAI Post ID: 2078718154289848621 Text: 🚨 The Chinese AI lab that pioneered million-token context windows just launched their own Claude Code. Free. Open sou
X post: User: @Mr_Salio Post ID: 2073763635680182772 Text: Gemini 3.5 Pro Leak: It beats Mythos 👀 >Targeting a July 17 release >Massive 2M token context window >New "Deep Think"
X post: User: @Mr_Salio Post ID: 2073763635680182772 Text: Gemini 3.5 Pro Leak: It beats Mythos 👀 >Targeting a July 17 release >Massive 2M token context window >New "Deep Think"
X post: User: @Mr_Salio Post ID: 2073763635680182772 Text: Gemini 3.5 Pro Leak: It beats Mythos 👀 >Targeting a July 17 release >Massive 2M token context window >New "Deep Think"
X post: User: @Alan_Earn Post ID: 2073625466267975711 Text: A new Gemini 3.5 Pro leak just surfaced here what's rumored: - Expected release on July 17 - 2M token context window -
X post: User: @Alan_Earn Post ID: 2073625466267975711 Text: A new Gemini 3.5 Pro leak just surfaced here what's rumored: - Expected release on July 17 - 2M token context window -
X post: User: @Alan_Earn Post ID: 2073625466267975711 Text: A new Gemini 3.5 Pro leak just surfaced here what's rumored: - Expected release on July 17 - 2M token context window -
X post: User: @shmidtqq Post ID: 2071950221764948023 Text: I MADE MY AI AGENT 10X FASTER WITHOUT CHANGING THE MODEL not a smarter model, not a bigger context window, not another c
X post: User: @shmidtqq Post ID: 2071950221764948023 Text: I MADE MY AI AGENT 10X FASTER WITHOUT CHANGING THE MODEL not a smarter model, not a bigger context window, not another c
X post: User: @shmidtqq Post ID: 2071950221764948023 Text: I MADE MY AI AGENT 10X FASTER WITHOUT CHANGING THE MODEL not a smarter model, not a bigger context window, not another c
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @riyazmd774 Post ID: 2067294202355208688 Text: CHINA JUST DROPPED AN AI CODING MODEL WITH A 1M CONTEXT WINDOW. And I connected it to Claude Code to see what it could
X post: User: @JackWoth98 Post ID: 2066593613098975384 Text: Custom Statusline in Antigravity CLI🚦⚙️ Track agent state, context window, subagents all right from the footer of you
X post: User: @JackWoth98 Post ID: 2066593613098975384 Text: Custom Statusline in Antigravity CLI🚦⚙️ Track agent state, context window, subagents all right from the footer of you
X post: User: @JackWoth98 Post ID: 2066593613098975384 Text: Custom Statusline in Antigravity CLI🚦⚙️ Track agent state, context window, subagents all right from the footer of you
AutoGPT's AutoPilot now supports a local transport (`CHAT_USE_LOCAL`) that routes chat to a self-hosted OpenAI-compatible endpoint (e.g., Ollama, vLLM) instead of requiring an OpenRouter or Anthropic API key, with runtime context-window detection and compaction fixes to prevent prompt corruption.
AutoGPT's AutoPilot now supports a local transport (`CHAT_USE_LOCAL`) that routes chat to a self-hosted OpenAI-compatible endpoint (e.g., Ollama, vLLM) instead of requiring an OpenRouter or Anthropic API key, with runtime context-window detection and compaction fixes to prevent prompt corruption.
AutoGPT's AutoPilot now supports a local transport (`CHAT_USE_LOCAL`) that routes chat to a self-hosted OpenAI-compatible endpoint (e.g., Ollama, vLLM) instead of requiring an OpenRouter or Anthropic API key, with runtime context-window detection and compaction fixes to prevent prompt corruption.
AnythingLLM's agent summarizer tool now walks document content one chunk at a time instead of relying on a LangChain map-reduce chain that failed to converge with small context window models. The new approach chunks at 45% of the model's context window, extracts key points per section while carrying forward prior summaries, and asks the user before continuing past the first three sections.
AnythingLLM's agent summarizer tool now walks document content one chunk at a time instead of relying on a LangChain map-reduce chain that failed to converge with small context window models. The new approach chunks at 45% of the model's context window, extracts key points per section while carrying forward prior summaries, and asks the user before continuing past the first three sections.
AnythingLLM's agent summarizer tool now walks document content one chunk at a time instead of relying on a LangChain map-reduce chain that failed to converge with small context window models. The new approach chunks at 45% of the model's context window, extracts key points per section while carrying forward prior summaries, and asks the user before continuing past the first three sections.
When an agent uses web-scraping or document-summarizer skills and content exceeds the context window, the summarization fallback now correctly creates LangChain chat models for Azure OpenAI and Dell Pro AI Studio providers, preventing the agent from crashing with an 'Unsupported provider' error.
When an agent uses web-scraping or document-summarizer skills and content exceeds the context window, the summarization fallback now correctly creates LangChain chat models for Azure OpenAI and Dell Pro AI Studio providers, preventing the agent from crashing with an 'Unsupported provider' error.
When an agent uses web-scraping or document-summarizer skills and content exceeds the context window, the summarization fallback now correctly creates LangChain chat models for Azure OpenAI and Dell Pro AI Studio providers, preventing the agent from crashing with an 'Unsupported provider' error.
Goose introduced an unstable ACP method that allows runtime replacement of the base system prompt and appending of keyed additional instructions per session, shifting prompt control from static configuration to dynamic agent-driven updates.
Goose introduced an unstable ACP method that allows runtime replacement of the base system prompt and appending of keyed additional instructions per session, shifting prompt control from static configuration to dynamic agent-driven updates.
Goose introduced an unstable ACP method that allows runtime replacement of the base system prompt and appending of keyed additional instructions per session, shifting prompt control from static configuration to dynamic agent-driven updates.
AnythingLLM agent tools previously passed an object instead of a string to the context limit caller, causing a silent fallback to an 8K context window. This update adds type validation and error logging to ensure user-defined context windows are respected, preventing errors when processing large documents.
AnythingLLM agent tools previously passed an object instead of a string to the context limit caller, causing a silent fallback to an 8K context window. This update adds type validation and error logging to ensure user-defined context windows are respected, preventing errors when processing large documents.
AnythingLLM agent tools previously passed an object instead of a string to the context limit caller, causing a silent fallback to an 8K context window. This update adds type validation and error logging to ensure user-defined context windows are respected, preventing errors when processing large documents.
Added first-class support for Cerebras as an OpenAI-compatible LLM provider in anything-llm, enabling chat and agent workflows to use Cerebras models via existing infrastructure with minimal configuration changes.
Added first-class support for Cerebras as an OpenAI-compatible LLM provider in anything-llm, enabling chat and agent workflows to use Cerebras models via existing infrastructure with minimal configuration changes.
Added first-class support for Cerebras as an OpenAI-compatible LLM provider in anything-llm, enabling chat and agent workflows to use Cerebras models via existing infrastructure with minimal configuration changes.
The agent runner's cold-start history reconstruction previously filtered only text content, silently discarding thinking, tool_use, and tool_result blocks. This caused multi-turn reasoning continuity loss and API 400 errors when sessions restarted or switched directories. The fix introduces a block-aware serializer that preserves these structured blocks during session replay, restoring accurate tool-call execution and model reasoning traces.
The agent runner's cold-start history reconstruction previously filtered only text content, silently discarding thinking, tool_use, and tool_result blocks. This caused multi-turn reasoning continuity loss and API 400 errors when sessions restarted or switched directories. The fix introduces a block-aware serializer that preserves these structured blocks during session replay, restoring accurate tool-call execution and model reasoning traces.
The agent runner's cold-start history reconstruction previously filtered only text content, silently discarding thinking, tool_use, and tool_result blocks. This caused multi-turn reasoning continuity loss and API 400 errors when sessions restarted or switched directories. The fix introduces a block-aware serializer that preserves these structured blocks during session replay, restoring accurate tool-call execution and model reasoning traces.
Introduces a SessionStart hook that automatically loads the vault's _CLAUDE.md configuration into the AI assistant's context at launch, removing the need to re-read the file on each subsequent command.
Introduces a SessionStart hook that automatically loads the vault's _CLAUDE.md configuration into the AI assistant's context at launch, removing the need to re-read the file on each subsequent command.
Introduces a SessionStart hook that automatically loads the vault's _CLAUDE.md configuration into the AI assistant's context at launch, removing the need to re-read the file on each subsequent command.