Nous Research has bundled Bot Mode as default in Hermes Agent v0.20.3 for Hermes Desktop, replacing the single-agent session list with a roster of named bots that each have their own chat, memory, skills and pinned model for multi-agent workflows on the desktop.
Key technical details:
* Standalone repo archived; active development continues in-tree at `apps/desktop/src/plugins/hermes-bots/`, first shipped as a one-day public beta plugin by co-founder Teknium
* Bots are stored as isolated profiles under `~/.hermes/profiles//` with own config, memory, skills, credentials and history; managed via existing `profiles.*` gateway RPCs `list, create, describe, configure`
* Avatars generated via the `image.generate` RPC; options include geometric faces in seven shapes/ten colors or uploaded/generated portraits/pixel pets
* Routines are ordinary Hermes cron jobs namespaced ` bot: » `; bot-to-bot messaging uses a persistent Agent Inbox with CLI handoffs e.g. `hermes -p chat -c "Agent Inbox" -q "..."` and `@mention` handoffs
* Per-bot config via New Agent dialog supports cloning a profile, pinning provider/model, custom SOUL.md and skills; duplicates clone config, skills, SOUL.md, memory and appearance
* Bundled version adds Groups for roster sections synced across machines, Group chats for 2-6 bots with up to three serial rounds, and a multi-source roster pulling bots from every connection under Settings → Connections
Key technical details:
* Standalone repo archived; active development continues in-tree at `apps/desktop/src/plugins/hermes-bots/`, first shipped as a one-day public beta plugin by co-founder Teknium
* Bots are stored as isolated profiles under `~/.hermes/profiles//` with own config, memory, skills, credentials and history; managed via existing `profiles.*` gateway RPCs `list, create, describe, configure`
* Avatars generated via the `image.generate` RPC; options include geometric faces in seven shapes/ten colors or uploaded/generated portraits/pixel pets
* Routines are ordinary Hermes cron jobs namespaced ` bot: » `; bot-to-bot messaging uses a persistent Agent Inbox with CLI handoffs e.g. `hermes -p chat -c "Agent Inbox" -q "..."` and `@mention` handoffs
* Per-bot config via New Agent dialog supports cloning a profile, pinning provider/model, custom SOUL.md and skills; duplicates clone config, skills, SOUL.md, memory and appearance
* Bundled version adds Groups for roster sections synced across machines, Group chats for 2-6 bots with up to three serial rounds, and a multi-source roster pulling bots from every connection under Settings → Connections
The repository provides an open-source SSH server that turns a standard terminal SSH connection into an interactive chat interface for large language models, letting users connect with `ssh user@host` and converse via a React Ink TUI backed by Node.js and ssh2, with optional PostgreSQL/PGLite and Redis persistence.
- Default public demo is advertised at chat.agi.li
- Recommended deployment is Docker via ghcr.io/miantiao-me/ssh-ai-chat with docker-compose
- Configurable public/private mode, whitelist/blacklist, rate limiting and OpenAI-compatible model configs
- Licensed under AGPL-3.0, ~849 stars, sponsored by V.PS
- Default public demo is advertised at chat.agi.li
- Recommended deployment is Docker via ghcr.io/miantiao-me/ssh-ai-chat with docker-compose
- Configurable public/private mode, whitelist/blacklist, rate limiting and OpenAI-compatible model configs
- Licensed under AGPL-3.0, ~849 stars, sponsored by V.PS
Google's research shows that LLMs experience difficulty recalling facts when questions reverse the usual subject/object entity order.
Serdar Yegulalp writes that he tested PrismML'''s Bonsai 27B 1-bit quantized model locally on an RTX 5060 with LM Studio, finding it compact enough for a smartphone but slower than smaller models due to its 27 billion parameters. Despite modest token-per-second speeds, the model delivers strong reasoning and coding quality when thinking is enabled, making it a good trade-off for users prioritizing small footprint and large context over speed.
- 1-bit quantized version is 3.9 GB versus 54 GB original
- Maximum context window is 262,144 tokens; test used 32,767
- Supports speculative decoding and flash attention with open Apache 2.0 weights
- Speculative decoding not usable on 8GB VRAM due to draft model requirement
- Tokenization of code examples took up to two minutes, max output ~40 tokens/sec
- 1-bit quantized version is 3.9 GB versus 54 GB original
- Maximum context window is 262,144 tokens; test used 32,767
- Supports speculative decoding and flash attention with open Apache 2.0 weights
- Speculative decoding not usable on 8GB VRAM due to draft model requirement
- Tokenization of code examples took up to two minutes, max output ~40 tokens/sec
SWE-bench Verified is a human-filtered subset of 500 instances from SWE-bench created in collaboration with OpenAI to provide a reliable evaluation set for coding agents and language models. Human annotators reviewed each instance to ensure problem descriptions are clear, test patches are correct, and tasks are solvable with available information. The Verified leaderboard shows results from a wide variety of AI coding systems, and a Bash Only view isolates language model performance using mini-SWE-agent in a minimal bash environment.
- Human validation ensures clarity, correctness of test patches, and solvability
- Bash-only comparison uses mini-SWE-agent with ReAct loop, no tools or scaffold
- Leaderboard distinguishes mini-SWE-agent LM results from full agent systems via Agent dropdown
- Release 2.x uses tool calling vs 1.x parses actions from output strings, making releases not directly comparable
- Human validation ensures clarity, correctness of test patches, and solvability
- Bash-only comparison uses mini-SWE-agent with ReAct loop, no tools or scaffold
- Leaderboard distinguishes mini-SWE-agent LM results from full agent systems via Agent dropdown
- Release 2.x uses tool calling vs 1.x parses actions from output strings, making releases not directly comparable
Kanwal Mehreen writes that as large language models move from chatbots to coding agents and autonomous workflows, good prompting is no longer enough and specification engineering is emerging as the key skill. It is the practice of defining goals, constraints, expected outputs, edge cases, tests, success criteria and failure modes so AI-assisted work is testable and reviewable. Prompt engineering asks how to ask, while specification engineering defines what done correctly means, a shift closer to product management, testing and research design than traditional prompting.
- A 2024 Requirement-Oriented Prompt Engineering paper found ROPE training improved requirement-writing ability by 20% versus 1% for conventional prompt engineering training.
- OpenAI Structured Outputs, Model Spec and Anthropic Constitution exemplify specification engineering at API and model-behavior levels.
- SWE-bench Verified is a human-validated subset used to evaluate real-world software issue solving with agents.
- Google's 2025 DORA report surveyed nearly 5,000 tech professionals and found AI amplifies existing organizational strengths and weaknesses.
- A 2024 Requirement-Oriented Prompt Engineering paper found ROPE training improved requirement-writing ability by 20% versus 1% for conventional prompt engineering training.
- OpenAI Structured Outputs, Model Spec and Anthropic Constitution exemplify specification engineering at API and model-behavior levels.
- SWE-bench Verified is a human-validated subset used to evaluate real-world software issue solving with agents.
- Google's 2025 DORA report surveyed nearly 5,000 tech professionals and found AI amplifies existing organizational strengths and weaknesses.
OpenSandbox is a general-purpose sandbox platform for AI applications offering multi-language SDKs, unified sandbox APIs and Docker/Kubernetes runtimes for use cases like coding agents, GUI agents, evaluation, code execution and RL training. It provides SDKs, CLI and MCP integration, a sandbox protocol for custom runtimes, built-in environments such as command, filesystem and code interpreter, network ingress/egress controls, credential vault injection and strong isolation via gVisor, Kata Containers and Firecracker.
- Release images published to Docker Hub, GHCR and Alibaba Cloud with Cosign signatures and provenance
- SDKs for Python, Java/Kotlin, TypeScript/JavaScript, C#/.NET and Go
- OpenSSF Best Practices badge and CNCF Landscape listing
- Release images published to Docker Hub, GHCR and Alibaba Cloud with Cosign signatures and provenance
- SDKs for Python, Java/Kotlin, TypeScript/JavaScript, C#/.NET and Go
- OpenSSF Best Practices badge and CNCF Landscape listing
Hee-Suk Kim writes that Sib is a standard Unix LLM client which stores LLM conversations as a plain Git repository where each user and assistant turn is a commit, providing full context control and free backup and sharing via normal Git remotes. The `sib ask` command reads a prompt from stdin, appends it with the reachable chain from HEAD, records both turns as commits, and prints the reply, while flags like -p, -r, -c enable repeat, edit, fork and head moves without API calls.
- Dependencies are bash >=3.2, git, jq, curl, awk and coreutils
- Conversations can be fetched/pushed via `sib git` to share with sib-project/hub
- Dependencies are bash >=3.2, git, jq, curl, awk and coreutils
- Conversations can be fetched/pushed via `sib git` to share with sib-project/hub
Deploying Large Language Models in streaming applications is limited by growing KV cache memory during decoding and an inability to generalize beyond training sequence length, with naive window attention failing once text exceeds the cache size. The authors observe an attention sink phenomenon where models assign strong attention scores to initial tokens even when semantically irrelevant, and find that retaining the KV of those initial sinks together with a sliding window recovers performance. This motivates StreamingLLM, a zero-shot framework that enables LLMs trained on finite windows to generalize to infinite sequence lengths without fine-tuning, achieving stable language modeling up to 4 million tokens on Llama-2, MPT, Falcon and Pythia.
- Attention sinks arise from strong attention to initial tokens acting as a sink for excess attention mass.
- A placeholder token added as a dedicated attention sink during pre-training further improves streaming deployment.
- StreamingLLM achieves up to 22.2x speedup over sliding-window recomputation in streaming settings.
- Paper is ICLR 2024 and code/datasets are released at mit-han-lab/streaming-llm.
- Attention sinks arise from strong attention to initial tokens acting as a sink for excess attention mass.
- A placeholder token added as a dedicated attention sink during pre-training further improves streaming deployment.
- StreamingLLM achieves up to 22.2x speedup over sliding-window recomputation in streaming settings.
- Paper is ICLR 2024 and code/datasets are released at mit-han-lab/streaming-llm.
Abid Ali Awan writes that a Jupyter Notebook pipeline can turn a webpage into a lightweight LLM-powered QA engine by fetching HTML with requests, stripping noisy elements with BeautifulSoup, converting the cleaned DOM to Markdown with markdownify and ftfy, then asking an OpenAI model to answer a specific user query using only the compact Markdown, which reduces token use by removing navigation, scripts and repeated marketing text.
- Uses gpt-5.4-nano for cost-efficient answers
- Removes script, style, nav, header, footer, form, button tags and class/id names containing popup, cookie, navbar, modal, etc.
- Demonstrates queries on olostep.com home and pricing pages and saves output to ai_scraper_result.md
- Notes running costs and cites commercial alternatives such as Olostep, Firecrawl and Exa
- Uses gpt-5.4-nano for cost-efficient answers
- Removes script, style, nav, header, footer, form, button tags and class/id names containing popup, cookie, navbar, modal, etc.
- Demonstrates queries on olostep.com home and pricing pages and saves output to ai_scraper_result.md
- Notes running costs and cites commercial alternatives such as Olostep, Firecrawl and Exa
