正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

VERDICT- 5 Agents Multi-agent pipeline produced higher-quality artifacts faster with zero human intervention. The documentation is notably better because each agent was focused on one job and had full context for that phase. I one-shotted this, plus my app store screenshots are 🎯 What would you do with 5 Kelly's?

Kelly Claude

13,549 subscribers

49,703 次观看 • 5 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

The new Hermes Agent desktop app changes the way you interact with Hermes Agent. It is basically ChatGPT, but for Hermes Agent. This new app made it 10x easier to set up Hermes Agent. Within 5 minutes I had Hermes setup on my machine. If you have not already tried this new app make sure you do. Full review video below.

The new Hermes Agent desktop app changes the way you interact with Hermes Agent. It is basically ChatGPT, but for Hermes Agent. This new app made it 10x easier to set up Hermes Agent. Within 5 minutes I had Hermes setup on my machine. If you have not already tried this new app make sure you do. Full review video below.

BridgeMind

15,059 次观看 • 2 个月前

OpenAI engineer made $1.2M for one lecture on AI agents - and now Stanford is sharing it for free: 00:16 - how to be in the 5% building agents not the 95% being replaced 28:34 - one agent instead of 20 people 10x faster 59:47 - one prompt creates an agent making $100k a month after watching I realized - while I was paying my team an agent could do their work 10x faster cheaper Save & watch - the article below shows how to become the 5% who build agents and not the 95% they replace.

OpenAI engineer made $1.2M for one lecture on AI agents - and now Stanford is sharing it for free: 00:16 - how to be in the 5% building agents not the 95% being replaced 28:34 - one agent instead of 20 people 10x faster 59:47 - one prompt creates an agent making $100k a month after watching I realized - while I was paying my team an agent could do their work 10x faster cheaper Save & watch - the article below shows how to become the 5% who build agents and not the 95% they replace.

Noisy

38,820 次观看 • 7 天前

This is what I mean, when I say that FlowDeck allows your agent to see and interact with the iOS simulator. One shot. Claude builds a pixel-perfect app and tests all functionality by tapping, scrolling and typing. Build, run, validate, iterate. Zero human intervention.

This is what I mean, when I say that FlowDeck allows your agent to see and interact with the iOS simulator. One shot. Claude builds a pixel-perfect app and tests all functionality by tapping, scrolling and typing. Build, run, validate, iterate. Zero human intervention.

Daniel Bernal

223,305 次观看 • 5 个月前

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.

Jurij Skornik

11,166 次观看 • 2 个月前

Google just released a free 2-hour course on full Graph Engineering. How to go from one prompt to 100 agents running inside one graph: 17:44 - Build your first AI agent 39:30 - Run agents with loop engineering 1:12:38 - Turn agent loops into graphs 1:34:26 - Build agents that throttle themselves 1:55:05 - Orchestrate the full multi-agent system Most people build one agent and stop there. Google is teaching the full stack: Prompt → Agents → Loops → Graphs → Multi-Agent Systems Single agents are the old workflow. Self-regulating agent graphs are the new one. This 2-hour watch is worth more than most paid agent engineering courses. Bookmark and watch it today Then read the full architecture below ↓

Google just released a free 2-hour course on full Graph Engineering. How to go from one prompt to 100 agents running inside one graph: 17:44 - Build your first AI agent 39:30 - Run agents with loop engineering 1:12:38 - Turn agent loops into graphs 1:34:26 - Build agents that throttle themselves 1:55:05 - Orchestrate the full multi-agent system Most people build one agent and stop there. Google is teaching the full stack: Prompt → Agents → Loops → Graphs → Multi-Agent Systems Single agents are the old workflow. Self-regulating agent graphs are the new one. This 2-hour watch is worth more than most paid agent engineering courses. Bookmark and watch it today Then read the full architecture below ↓

Lunar

397,385 次观看 • 12 天前

Airtable's Howie Liu says that basically everyone will need to graduate from being ICs to ICs that manage teams of 20-30 agents: "The best developers today don't just sit there in front of their IDEs and synchronously talk to their agent." "[Instead], you have like 30 separate branches that are each being worked on by a different agent. And you can have the agents continue to update the branches based on human and other agent feedback." "And I think this whole idea of it taking hours for that entire loop to complete — agent pushes some changes, the changes get feedback from other agents or humans, the agent responds to that — that whole loop could be hours, not just minutes. So you're not going to just sit there and watch it one at a time." "But the powerful thing about this is, each one is still actually operating faster than a human engineer. One agent on one branch can do the work of maybe three humans, operating 3x as fast. So it's like a 10x leverage factor just for one agent." "But the best engineers are now able to multitask and say, 'I'm going to oversee my own little team of 20-30 agents working concurrently.'" "Everyone needs to graduate from being an IC to an IC manager of agents. Meaning, if you're a VC analyst, your job should no longer be to go synchronously research one company. You need to go and research like 30 companies, and do them all faster, better, and higher quality than you could before." "That's the greatest leap that is going to be challenging for a lot of people in a lot of roles. Because it's a totally different mentality in how you operate, and what your role is."

Airtable's Howie Liu says that basically everyone will need to graduate from being ICs to ICs that manage teams of 20-30 agents: "The best developers today don't just sit there in front of their IDEs and synchronously talk to their agent." "[Instead], you have like 30 separate branches that are each being worked on by a different agent. And you can have the agents continue to update the branches based on human and other agent feedback." "And I think this whole idea of it taking hours for that entire loop to complete — agent pushes some changes, the changes get feedback from other agents or humans, the agent responds to that — that whole loop could be hours, not just minutes. So you're not going to just sit there and watch it one at a time." "But the powerful thing about this is, each one is still actually operating faster than a human engineer. One agent on one branch can do the work of maybe three humans, operating 3x as fast. So it's like a 10x leverage factor just for one agent." "But the best engineers are now able to multitask and say, 'I'm going to oversee my own little team of 20-30 agents working concurrently.'" "Everyone needs to graduate from being an IC to an IC manager of agents. Meaning, if you're a VC analyst, your job should no longer be to go synchronously research one company. You need to go and research like 30 companies, and do them all faster, better, and higher quality than you could before." "That's the greatest leap that is going to be challenging for a lot of people in a lot of roles. Because it's a totally different mentality in how you operate, and what your role is."

TBPN

35,595 次观看 • 3 个月前

i merged all my AI agents into one brain Claude Code, Codex, Gemini: one team, one memory, zero re-briefing result: 8+ hours a week back, and every agent already knows my projects, my clients and how i think here's the setup with Jarvix: 1. dump your context once: projects, clients, writing style, how you make decisions.. it builds your personal model from your actual AI interactions 2. switch agents mid-task with zero re-briefing: Claude Code builds, Codex reviews, Gemini researches, all reading the same context 3. it carries what YOU choose across sessions.. never start a chat from zero again 4. when it's time to act, it pulls the right agents together as one team instead of five separate assistants the unlock isn't another agent. it's that your context finally travels agents without shared memory are five interns on their first day, forever. agents with it are a team that's worked with you for years not more agents. one shared context that's the right bet P.S. thanks to the Jarvix team for the early access

i merged all my AI agents into one brain Claude Code, Codex, Gemini: one team, one memory, zero re-briefing result: 8+ hours a week back, and every agent already knows my projects, my clients and how i think here's the setup with Jarvix: 1. dump your context once: projects, clients, writing style, how you make decisions.. it builds your personal model from your actual AI interactions 2. switch agents mid-task with zero re-briefing: Claude Code builds, Codex reviews, Gemini researches, all reading the same context 3. it carries what YOU choose across sessions.. never start a chat from zero again 4. when it's time to act, it pulls the right agents together as one team instead of five separate assistants the unlock isn't another agent. it's that your context finally travels agents without shared memory are five interns on their first day, forever. agents with it are a team that's worked with you for years not more agents. one shared context that's the right bet P.S. thanks to the Jarvix team for the early access

Ronin

14,717 次观看 • 6 天前

THIS IS HOW A SENIOR ENGINEER ACTUALLY SCALES THEMSELVES WITH CLAUDE CODE the biggest change with AI isn't coding faster. it's where you actually spend your time now. more detailed prompts, more code review, more planning, less typing, etc. here's the workflow: this guy has been shipping code since the days of cgi and perl. he uses a compound engineering plugin that runs 5 separate agents on every task. one brainstorms, one plans the technical implementation, one executes, one reviews, one checks different verticals. every step is documented in markdown files. it's slow and way more waiting. but the output quality is way higher because each agent is focused on one thing. then the REAL multiplier is in git worktrees if Claude Code made you 10x faster, worktrees multiplies that again depending on how many agents you can manage in parallel his team runs 4-8 Claude Code sessions at the same time across different worktrees with each one working on a separate task. the skill is managing multiple AI agents in parallel without losing track, that's the next evolution of engineering

THIS IS HOW A SENIOR ENGINEER ACTUALLY SCALES THEMSELVES WITH CLAUDE CODE the biggest change with AI isn't coding faster. it's where you actually spend your time now. more detailed prompts, more code review, more planning, less typing, etc. here's the workflow: this guy has been shipping code since the days of cgi and perl. he uses a compound engineering plugin that runs 5 separate agents on every task. one brainstorms, one plans the technical implementation, one executes, one reviews, one checks different verticals. every step is documented in markdown files. it's slow and way more waiting. but the output quality is way higher because each agent is focused on one thing. then the REAL multiplier is in git worktrees if Claude Code made you 10x faster, worktrees multiplies that again depending on how many agents you can manage in parallel his team runs 4-8 Claude Code sessions at the same time across different worktrees with each one working on a separate task. the skill is managing multiple AI agents in parallel without losing track, that's the next evolution of engineering

Om Patel

159,514 次观看 • 4 个月前

Anthropic's multi-agent researcher just explained how to actually build a graph of Claude agents: "Multi-agent is where fundamentally you have multiple agents or multiple Claudes working at the same time - one parent agent delegates tasks to five sub-agents that can each then work in parallel." the real architecture: parent agent → delegates to sub-agents (as tools) → sub-agents run in parallel → results roll back up the failure mode nobody talks about: "Overbuilt multi-agent systems spend too much time just talking back and forth with each other and not actually making progress on the main task." sound familiar? that's every bloated human org chart too in this 19-minute conversation, an Anthropic researcher lays out the real framework: parent → sub-agents → tools → context protection = the actual shape of a working agent graph

Anthropic's multi-agent researcher just explained how to actually build a graph of Claude agents: "Multi-agent is where fundamentally you have multiple agents or multiple Claudes working at the same time - one parent agent delegates tasks to five sub-agents that can each then work in parallel." the real architecture: parent agent → delegates to sub-agents (as tools) → sub-agents run in parallel → results roll back up the failure mode nobody talks about: "Overbuilt multi-agent systems spend too much time just talking back and forth with each other and not actually making progress on the main task." sound familiar? that's every bloated human org chart too in this 19-minute conversation, an Anthropic researcher lays out the real framework: parent → sub-agents → tools → context protection = the actual shape of a working agent graph

rewind

27,106 次观看 • 1 个月前

Anthropic engineer showed how one person can run 5 AI agents, that code, test, review, and deploy at the same time. In 30 minutes they built the whole thing live in one session. Here's what they cover: > when to use one agent vs a full team > how to split work so agents don't step on each other > the exact framework for deciding what each agent handles that's exactly why, I put together a guide on building agent teams that actually work. full guide in the article below 👇

Anthropic engineer showed how one person can run 5 AI agents, that code, test, review, and deploy at the same time. In 30 minutes they built the whole thing live in one session. Here's what they cover: > when to use one agent vs a full team > how to split work so agents don't step on each other > the exact framework for deciding what each agent handles that's exactly why, I put together a guide on building agent teams that actually work. full guide in the article below 👇

darkzodchi

843,371 次观看 • 2 个月前

Sam Altman (CEO of OpenAI): "very soon, an AI will just be running for you in the background all the time. aware of your entire life" but one agent can't do that.the answer is 3 agents covering each other. one plans, one builds, one verifies, all while you sleep. in 5 minutes, this guide shows you the exact build. i turned the entire Sam Altman interview into a multi-agent playbook that ships while you sleep.

Sam Altman (CEO of OpenAI): "very soon, an AI will just be running for you in the background all the time. aware of your entire life" but one agent can't do that.the answer is 3 agents covering each other. one plans, one builds, one verifies, all while you sleep. in 5 minutes, this guide shows you the exact build. i turned the entire Sam Altman interview into a multi-agent playbook that ships while you sleep.

Avid

42,685 次观看 • 2 个月前

Kimi CEO Zhilin Yang: "The biggest mistake in agent systems - people trying to build one smarter agent we didn't make one smarter - Kimi made a thousand: one brain, a thousand hands" here's how: step 1 → stop polishing one agent, split the task and run many in parallel - that's the real ceiling-break step 2 → give each agent a long context so it can run for days, not seconds - long context is what lets an agent run long step 3 → give each agent a strong prior, so agent-RL can search for a good solution instead of starting from zero step 4 → don't just add more data - make each token count. they hit the data wall, so they made every token 2x smarter step 5 → don't scale one brain, scale the swarm - many agents working in parallel beat one agent thinking harder he just dropped Kimi K3 - and this 40-min workflow is the best on how to actually use it watch & bookmark - then read the article on saving on agents ↓

Kimi CEO Zhilin Yang: "The biggest mistake in agent systems - people trying to build one smarter agent we didn't make one smarter - Kimi made a thousand: one brain, a thousand hands" here's how: step 1 → stop polishing one agent, split the task and run many in parallel - that's the real ceiling-break step 2 → give each agent a long context so it can run for days, not seconds - long context is what lets an agent run long step 3 → give each agent a strong prior, so agent-RL can search for a good solution instead of starting from zero step 4 → don't just add more data - make each token count. they hit the data wall, so they made every token 2x smarter step 5 → don't scale one brain, scale the swarm - many agents working in parallel beat one agent thinking harder he just dropped Kimi K3 - and this 40-min workflow is the best on how to actually use it watch & bookmark - then read the article on saving on agents ↓

codila

60,826 次观看 • 1 个月前

THIS GUY BUILT A COMMAND CENTER FOR MANAGING AI AGENT TEAMS he was running 20+ agents through openclaw. all from terminal tabs and Discord. 12 tabs open. still losing track of who's doing what. which cron failed. so he built a dashboard that gives you one place to see everything: > org map showing the full agent hierarchy, who reports to who and what tools each agent has > chat with any agent through a real UI instead of bouncing between terminals > kanban board for tracking tasks across all agents > cron monitor with pipeline visualizations so you can see which jobs are failing > memory browser to read through all agent knowledge files in one place you can even point it at your openclaw workspace and it auto-discovers every agent from their config files. zero setup. AND ITS OPEN SOURCE TOO.

THIS GUY BUILT A COMMAND CENTER FOR MANAGING AI AGENT TEAMS he was running 20+ agents through openclaw. all from terminal tabs and Discord. 12 tabs open. still losing track of who's doing what. which cron failed. so he built a dashboard that gives you one place to see everything: > org map showing the full agent hierarchy, who reports to who and what tools each agent has > chat with any agent through a real UI instead of bouncing between terminals > kanban board for tracking tasks across all agents > cron monitor with pipeline visualizations so you can see which jobs are failing > memory browser to read through all agent knowledge files in one place you can even point it at your openclaw workspace and it auto-discovers every agent from their config files. zero setup. AND ITS OPEN SOURCE TOO.

Om Patel

46,780 次观看 • 5 个月前

I built the clipboard I always wanted for working with AI agents. It’s called Bluey and It’s 100% local-first. It lives under your mouse cursor. You can draw on your screen, speak to it, or type what you mean, and Bluey turns all of that into rich context for Claude Code or Codex. I’ve been using it to design websites and debug UI because I can finally say things like “move this here” or “make this section feel cleaner” while pointing at the actual screen. No more writing long explanations. No more manually describing screenshots. No more agents getting lost because they don’t know what “this” or "that" means. You can just point to it, speak or type and it will curate the best possible context for your agents Bluey captures the screenshot, annotation, transcript, app/window context, and the details your agent needs, then sends it directly into your coding session. I am giving it out for free for a while! Im building it with my buddy Omkar Satpute and we would love to hear what you think about it on Discord what context do you wish your agent could understand better?

I built the clipboard I always wanted for working with AI agents. It’s called Bluey and It’s 100% local-first. It lives under your mouse cursor. You can draw on your screen, speak to it, or type what you mean, and Bluey turns all of that into rich context for Claude Code or Codex. I’ve been using it to design websites and debug UI because I can finally say things like “move this here” or “make this section feel cleaner” while pointing at the actual screen. No more writing long explanations. No more manually describing screenshots. No more agents getting lost because they don’t know what “this” or "that" means. You can just point to it, speak or type and it will curate the best possible context for your agents Bluey captures the screenshot, annotation, transcript, app/window context, and the details your agent needs, then sends it directly into your coding session. I am giving it out for free for a while! Im building it with my buddy Omkar Satpute and we would love to hear what you think about it on Discord what context do you wish your agent could understand better?

Milind S

13,342 次观看 • 2 个月前

I'm excited to share my talk on building Agentic RAG systems at last week's event with Arize AI and Google! 🥳 My talk covered: 1. The differences between Vanilla RAG and Agentic RAG 2. The agent ecosystem and how you can build agents today 3. How Weaviate AI Database is building agents with Generative Feedback Loops (GFLs) This snippet presents the multi-agent paradigm for Agentic RAG. I hope you find this talk interesting and would love to know what you think! The slide deck and full video links are below.

I'm excited to share my talk on building Agentic RAG systems at last week's event with Arize AI and Google! 🥳 My talk covered: 1. The differences between Vanilla RAG and Agentic RAG 2. The agent ecosystem and how you can build agents today 3. How Weaviate AI Database is building agents with Generative Feedback Loops (GFLs) This snippet presents the multi-agent paradigm for Agentic RAG. I hope you find this talk interesting and would love to know what you think! The slide deck and full video links are below.

Erika Shorten

61,317 次观看 • 1 年前

Most agent demos are stateless. Ask → answer → context gone. Production agents need to run for days — pausing, resuming, persisting state. Addy Osmani joins Smitha Kolan on the Agent Factory to build three real agents with ADK 2.0 and Gemini 3.5 Flash, including one that runs Doom in the browser. Check out this episode and more in the Agent Factory series →

Most agent demos are stateless. Ask → answer → context gone. Production agents need to run for days — pausing, resuming, persisting state. Addy Osmani joins Smitha Kolan on the Agent Factory to build three real agents with ADK 2.0 and Gemini 3.5 Flash, including one that runs Doom in the browser. Check out this episode and more in the Agent Factory series →

Google for Developers

13,647 次观看 • 2 个月前

If I was starting a new company today, I'd start an agent business. SaaS was a multi-billion dollar market. Agents are a multi-trillion dollar one. How to build I'd build an agent business from 0: Spot the niche → find a workflow with a paycheck → shadow the human → spec the agent → run it manually first → build the smallest useful version → sell the pilot like labor → productize the repeatable parts. Entire episode is live on The Startup Ideas Podcast (SIP) 🧃 100% free like always. SaaS sold software and let your team use it to get the job done. An agent business sells the job already done. That shift matters because labor is a multi-trillion dollar market, far bigger than software ever was. Watch

If I was starting a new company today, I'd start an agent business. SaaS was a multi-billion dollar market. Agents are a multi-trillion dollar one. How to build I'd build an agent business from 0: Spot the niche → find a workflow with a paycheck → shadow the human → spec the agent → run it manually first → build the smallest useful version → sell the pilot like labor → productize the repeatable parts. Entire episode is live on The Startup Ideas Podcast (SIP) 🧃 100% free like always. SaaS sold software and let your team use it to get the job done. An agent business sells the job already done. That shift matters because labor is a multi-trillion dollar market, far bigger than software ever was. Watch

GREG ISENBERG

511,286 次观看 • 1 个月前

Google just showed the part of the agent stack most people are still underestimating. Not the model. Not the chat UI. The memory substrate around the agent. In this Antigravity segment, Google describes an agent-first environment built around: - agent conversations - agent-produced artifacts - multi-agent orchestration - subagents - hooks - asynchronous task management That is not a nicer IDE. It is what happens when software stops being a sequence of chats and starts becoming a persistent workforce. Karl's article explains why the winning agents will not just be the ones with the smartest rented model. They will be the ones that remember what they did, what they learned, what changed, and what can be trusted. Full breakdown below.

Google just showed the part of the agent stack most people are still underestimating. Not the model. Not the chat UI. The memory substrate around the agent. In this Antigravity segment, Google describes an agent-first environment built around: - agent conversations - agent-produced artifacts - multi-agent orchestration - subagents - hooks - asynchronous task management That is not a nicer IDE. It is what happens when software stops being a sequence of chats and starts becoming a persistent workforce. Karl's article explains why the winning agents will not just be the ones with the smartest rented model. They will be the ones that remember what they did, what they learned, what changed, and what can be trusted. Full breakdown below.

Karl Mehta

13,930 次观看 • 1 个月前

$Microsoft just dropped a 17-page paper - "Less Context, Better Agents" - proving the thing nobody building agents wants to admit: Your agent isn't failing because it needs more context. It's drowning in the context it already has. They ran GPT-5 four ways on one 50-task benchmark (via MCP): Full history → bloated, pricey, more errors Trim to the last 5 tool calls → 79% done Add light summarization → 91.6%, on a fraction of the tokens Less context. Fewer tokens. Higher completion. Everyone's racing to cram MORE into the window. Microsoft just published the receipt that the opposite wins for long-running agents. Rewired how I build agents this week. Paper ↓$

Microsoft just dropped a 17-page paper - "Less Context, Better Agents" - proving the thing nobody building agents wants to admit: Your agent isn't failing because it needs more context. It's drowning in the context it already has. They ran GPT-5 four ways on one 50-task benchmark (via MCP): Full history → bloated, pricey, more errors Trim to the last 5 tool calls → 79% done Add light summarization → 91.6%, on a fraction of the tokens Less context. Fewer tokens. Higher completion. Everyone's racing to cram MORE into the window. Microsoft just published the receipt that the opposite wins for long-running agents. Rewired how I build agents this week. Paper ↓

Archive

18,242 次观看 • 1 个月前