正在加载视频...

视频加载失败

返回首页

Two new things just dropped in Kiro CLI: /goal — set a target, define max iterations, walk away. the agent doesn't stop until it has proof it's done. Ctrl+S — toggle between steering and queuing mid-session. inject prompts between turns instead of waiting for the LLM to finish. Agentic...

kirodotdev's profile picture

Kiro

21,105 subscribers

11,697 次观看 • 3 个月前 •via X (Twitter)

LIVE

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

“i wish i got more laps just to try things and test the setup in different ways. a lot of what i’ve been doing is just hoping it works. but we’ve done a good job as a team, so i have to be satisfied in the end that i at least got on the track today” 🙃
0:13

“i wish i got more laps just to try things and test the setup in different ways. a lot of what i’ve been doing is just hoping it works. but we’ve done a good job as a team, so i have to be satisfied in the end that i at least got on the track today” 🙃

ray

12,830 次观看 • 5 个月前

Sometimes a connection enters your field and it's unexplainable, you just remember them. There's no forcing this remembrance. You just know. Both can sit in silence and recognize each other like you've done this a thousand lifetimes before. It's the unspoken energy that forms between two souls that gets woven into the Akashic field, and when you meet again, there's no logic that can define what this is. It's just felt. It's also not about forcing or trying to make sense of it. It's about allowing what's real to move through you effortlessly. Because when the connection is real, you don't need proof. The proof is trusting what you feel. ~ Elizabethcopolla ✨🙌🏽💫
0:11

Sometimes a connection enters your field and it's unexplainable, you just remember them. There's no forcing this remembrance. You just know. Both can sit in silence and recognize each other like you've done this a thousand lifetimes before. It's the unspoken energy that forms between two souls that gets woven into the Akashic field, and when you meet again, there's no logic that can define what this is. It's just felt. It's also not about forcing or trying to make sense of it. It's about allowing what's real to move through you effortlessly. Because when the connection is real, you don't need proof. The proof is trusting what you feel. ~ Elizabethcopolla ✨🙌🏽💫

🧬Maxpein🧬

17,851 次观看 • 9 个月前

Double-Slit Experiment ✍️ It shows that tiny things like electrons act like ripples in water instead of solid marbles. When a particle is fired at two slits, it doesn't just go through one or the other; it travels as a wave of probability that passes through both at the same time. These two waves then overlap and interfere with each other, creating a pattern of light and dark stripes on a back wall. This shows that at the quantum level, reality isn't "set" until we actually measure it. It behaves as a spread-out wave of possibilities until the moment it hits a target. Video 📸 : umtiquinhodefisica
0:10

Double-Slit Experiment ✍️ It shows that tiny things like electrons act like ripples in water instead of solid marbles. When a particle is fired at two slits, it doesn't just go through one or the other; it travels as a wave of probability that passes through both at the same time. These two waves then overlap and interfere with each other, creating a pattern of light and dark stripes on a back wall. This shows that at the quantum level, reality isn't "set" until we actually measure it. It behaves as a spread-out wave of possibilities until the moment it hits a target. Video 📸 : umtiquinhodefisica

ScieVision

60,696 次观看 • 5 个月前

REDISTRICTING: For the first time in American history, a court has overturned a congressional map between a primary and a general. Missouri's Supreme Court is throwing out the results of the state's primary and the millions of dollars candidates spent campaigning using the state's new maps. In just two months, the state must figure out how to hold a primary, nominate new candidates, and run campaigns.
0:10

REDISTRICTING: For the first time in American history, a court has overturned a congressional map between a primary and a general. Missouri's Supreme Court is throwing out the results of the state's primary and the millions of dollars candidates spent campaigning using the state's new maps. In just two months, the state must figure out how to hold a primary, nominate new candidates, and run campaigns.

@amuse

20,795 次观看 • 16 天前

It's Friday, I'm love!😍 In PXL Wars, however, relationships can get toxic. 💔 Forge the 'In love' attribute and give your War PXL the power to 'seduce' its target, putting it out of action for a few turns. Just remember, love's a two-way street! #GameDev #PXLWars DEADPXLZ
0:10

It's Friday, I'm love!😍 In PXL Wars, however, relationships can get toxic. 💔 Forge the 'In love' attribute and give your War PXL the power to 'seduce' its target, putting it out of action for a few turns. Just remember, love's a two-way street! #GameDev #PXLWars DEADPXLZ

Adrian Fanatiu

11,140 次观看 • 2 年前

Being in this stadium tonight, it is so very clear how much of a bond has been built between this team and these fans. They’ve done their job, Up to Swiss federation to step up and make sure the moment lasts more than just a summer. Privilege to be here!
0:15

Being in this stadium tonight, it is so very clear how much of a bond has been built between this team and these fans. They’ve done their job, Up to Swiss federation to step up and make sure the moment lasts more than just a summer. Privilege to be here!

Girls on the Ball

47,290 次观看 • 1 年前

I'm seeing a lot of people saying it is just a bad upscaling instead of Gen AI, but it is indeed Gen AI animation. There are things going on in it that just don't make any sense for puppet animation. Here are some parts that make it clear, I slowed down the video so it's easier to notice the chunky parts. I checked the past ADs and it's possible to notice details (from an animator's POV) that are common and make sense for human work, so it was definitely a choice to use AI on this new AD
0:16

I'm seeing a lot of people saying it is just a bad upscaling instead of Gen AI, but it is indeed Gen AI animation. There are things going on in it that just don't make any sense for puppet animation. Here are some parts that make it clear, I slowed down the video so it's easier to notice the chunky parts. I checked the past ADs and it's possible to notice details (from an animator's POV) that are common and make sense for human work, so it was definitely a choice to use AI on this new AD

mei ✨ live2d animator

1,115,104 次观看 • 8 个月前

There’s two kinds of traders. The one that is able to walk away and the one that revenge trades. Overtrading kills more accounts than anything else. I used to be this way, getting angry that price wasn’t going my way and pushing the button again and again until I blew up all my accounts. It takes a lot of inner work, but believe me it’s worth it to work on not just your strategy and executions, but your mental game as well. Being able to walk away from the screen before it’s too late is a real skill set to work on. Don’t be the guy that revenge trades. Strategies that worked for me: 1. Writing down my game plan at the start of the morning. As long as I followed my plan, it’s a win. 2. Having pictures of family on my trading desk. Trading for others and not just yourself is important, it keeps you grounded. 3. Having alarms go off on my phone at the end of the session to bring me back to reality and remind me to walk away.
0:14

There’s two kinds of traders. The one that is able to walk away and the one that revenge trades. Overtrading kills more accounts than anything else. I used to be this way, getting angry that price wasn’t going my way and pushing the button again and again until I blew up all my accounts. It takes a lot of inner work, but believe me it’s worth it to work on not just your strategy and executions, but your mental game as well. Being able to walk away from the screen before it’s too late is a real skill set to work on. Don’t be the guy that revenge trades. Strategies that worked for me: 1. Writing down my game plan at the start of the morning. As long as I followed my plan, it’s a win. 2. Having pictures of family on my trading desk. Trading for others and not just yourself is important, it keeps you grounded. 3. Having alarms go off on my phone at the end of the session to bring me back to reality and remind me to walk away.

Tanja Trades

64,353 次观看 • 6 个月前

THIS JUST IN !! "One Eyed Bastard" (Live In Amsterdam) has arrived ⚡️ Thank you for making the song our 4th #1 at Alt radio from Saviors. Ya know, it's the first album to pull that off since '03, and we couldn't have done it without you. Now turn it upppp Nearly two years of Saviors and what a ride it's been. One more month of shows to send it off right. THANK YOU for being part of it!
0:13

THIS JUST IN !! "One Eyed Bastard" (Live In Amsterdam) has arrived ⚡️ Thank you for making the song our 4th #1 at Alt radio from Saviors. Ya know, it's the first album to pull that off since '03, and we couldn't have done it without you. Now turn it upppp Nearly two years of Saviors and what a ride it's been. One more month of shows to send it off right. THANK YOU for being part of it!

Green Day

42,733 次观看 • 1 年前

After a few more hours, I think I've figured out Opus 5. Opus 5 is trained to be more agentic than anything I've used. All Claude 5 models are like that. So what changes? The way to interact with Opus 5 or contextualize it won't work the same way as with other models. It loves exploring, so it doesn't need much guidance for it. Unique preferences, artifacts, and references compliment it well and enable cleaner and more effective exploration and execution. Now that it can explore more effectively on its own and understand intent better, the best thing to do is to get out of its way (e.g., it doesn't need examples of your preferences; a clear high-level description of it works best). It's truly agentic in that sense. A good first step to provide better context for Opus 5 is to distinguish between what's situational and what needs persistence. Regardless, persistent system prompts and CLAUDE.MD needs to stay lightweight. Remove memories and tool descriptions from these. CLAUDE.MD is also a great place to tap into progressive disclosure by linking command/skills to it. On the situational side, agent skills and auto-memory can leverage progressive disclosure and the improved ability of the model to use its external context/knowledge. Conflicting and unnecessary instructions, which are common at this layer (mainly to ensure reliability), are going to throw off this model easily. That's the biggest change I had to make. Simple, clean, and clear prompts and skills work best. I had to clean a lot of my skills and system prompts. The way I prompt remains the same (usually clear and well-scoped). MCP tool descriptions are also more descriptive and have been deduped from the system prompt. Anthropic released a guide on the new rules for context engineering, which was helpful here. I started to test the recommendations and created a little artifact with the things that worked along the way. This might feel like a lot of work. Believe me, it has been frustrating. But I think we can expect future frontier models to become more agentic and smarter at figuring out the right context/gaps. The best thing to do is to prepare for that now. Boris Cherny mentioned that Opus 5 is their least prompt-injectable model yet. I am not sure if that was something they intentionally trained for or if it emerged based on how it was trained, which is to be extremely agentic in nature and more direct in execution.
0:15

After a few more hours, I think I've figured out Opus 5. Opus 5 is trained to be more agentic than anything I've used. All Claude 5 models are like that. So what changes? The way to interact with Opus 5 or contextualize it won't work the same way as with other models. It loves exploring, so it doesn't need much guidance for it. Unique preferences, artifacts, and references compliment it well and enable cleaner and more effective exploration and execution. Now that it can explore more effectively on its own and understand intent better, the best thing to do is to get out of its way (e.g., it doesn't need examples of your preferences; a clear high-level description of it works best). It's truly agentic in that sense. A good first step to provide better context for Opus 5 is to distinguish between what's situational and what needs persistence. Regardless, persistent system prompts and CLAUDE.MD needs to stay lightweight. Remove memories and tool descriptions from these. CLAUDE.MD is also a great place to tap into progressive disclosure by linking command/skills to it. On the situational side, agent skills and auto-memory can leverage progressive disclosure and the improved ability of the model to use its external context/knowledge. Conflicting and unnecessary instructions, which are common at this layer (mainly to ensure reliability), are going to throw off this model easily. That's the biggest change I had to make. Simple, clean, and clear prompts and skills work best. I had to clean a lot of my skills and system prompts. The way I prompt remains the same (usually clear and well-scoped). MCP tool descriptions are also more descriptive and have been deduped from the system prompt. Anthropic released a guide on the new rules for context engineering, which was helpful here. I started to test the recommendations and created a little artifact with the things that worked along the way. This might feel like a lot of work. Believe me, it has been frustrating. But I think we can expect future frontier models to become more agentic and smarter at figuring out the right context/gaps. The best thing to do is to prepare for that now. Boris Cherny mentioned that Opus 5 is their least prompt-injectable model yet. I am not sure if that was something they intentionally trained for or if it emerged based on how it was trained, which is to be extremely agentic in nature and more direct in execution.

elvis

37,757 次观看 • 1 个月前

subagents are just recursive agents where you can apply different prompts + models depending on the task. since they’re just a primitive, Cursor cli can actually spawn subagents by calling cursor-agent in headless mode via shell commands. that’s what makes the cli so nice. you can extend it, experiment, and have a lot of fun exploring orchestration patterns. here’s one way to do it w. dynamic model selection: 1. create a subagents.mdc rule 2. drop in: ``` --- alwaysApply: true --- ALWAYS spawn subagents by running `cursor-agent -p [task] --output-format=text --force --model [model]` in the terminal. Each subagent should return a summary of the changes it made. Subagents should be used for ALL tasks You can adopt a fan-out pattern where you spawn subagents to perform parallel isolated tasks, and then fan-in the results. Use the following models: - `--model gpt-5` for reasoning, researching, and planning - `--model sonnet-4` for implementation ``` 3. start cursor cli and try it out you can also adjust the rule to be more explicit when it should use subagents, when not to, which models when etc.
0:15

subagents are just recursive agents where you can apply different prompts + models depending on the task. since they’re just a primitive, Cursor cli can actually spawn subagents by calling cursor-agent in headless mode via shell commands. that’s what makes the cli so nice. you can extend it, experiment, and have a lot of fun exploring orchestration patterns. here’s one way to do it w. dynamic model selection: 1. create a subagents.mdc rule 2. drop in: ``` --- alwaysApply: true --- ALWAYS spawn subagents by running `cursor-agent -p [task] --output-format=text --force --model [model]` in the terminal. Each subagent should return a summary of the changes it made. Subagents should be used for ALL tasks You can adopt a fan-out pattern where you spawn subagents to perform parallel isolated tasks, and then fan-in the results. Use the following models: - `--model gpt-5` for reasoning, researching, and planning - `--model sonnet-4` for implementation ``` 3. start cursor cli and try it out you can also adjust the rule to be more explicit when it should use subagents, when not to, which models when etc.

eric zakariasson

57,554 次观看 • 1 年前

Let me explain the agent loop, simple It's the core of every agentic system, and the part most people overcomplicate It's just this: 1. Send messages to the model 2. Model responds, maybe calls a tool 3. You run the tool 4. Append the result back to messages 5. Repeat until stop_reason is end_turn Step 4 is the whole thing, the write-back is what makes it an agent The model has to see what actually happened before it decides the next move That's the entire loop... understand this cold before you reach for a framework
0:14

Let me explain the agent loop, simple It's the core of every agentic system, and the part most people overcomplicate It's just this: 1. Send messages to the model 2. Model responds, maybe calls a tool 3. You run the tool 4. Append the result back to messages 5. Repeat until stop_reason is end_turn Step 4 is the whole thing, the write-back is what makes it an agent The model has to see what actually happened before it decides the next move That's the entire loop... understand this cold before you reach for a framework

Daniel San

12,514 次观看 • 3 个月前

Being a scientist is still the best job in the world. The constant frustrations, all the crap that goes with trying to keep the lab alive and move things forward - it all resets and washes away in those occasional moments when we catch a glimpse of nature revealing something remarkable and not seen before. This has been a good week - I saw a few completely amazing, wild new things; it will be a total slog to get them polished and out, but it doesn't matter because they've been seen and a new piece of the roadmap is revealed. Just feeling immensely grateful for the opportunity, and mentally sending a message back in time to childhood me (and anyone else who dreams of this now).
0:10

Being a scientist is still the best job in the world. The constant frustrations, all the crap that goes with trying to keep the lab alive and move things forward - it all resets and washes away in those occasional moments when we catch a glimpse of nature revealing something remarkable and not seen before. This has been a good week - I saw a few completely amazing, wild new things; it will be a total slog to get them polished and out, but it doesn't matter because they've been seen and a new piece of the roadmap is revealed. Just feeling immensely grateful for the opportunity, and mentally sending a message back in time to childhood me (and anyone else who dreams of this now).

Michael Levin

99,217 次观看 • 1 年前

Starmer resigning changes nothing. The mass migration, the rape gangs, the migrant crime, the green energy scam, these are Labour Party policy. Not just “Keir things.” So his replacement won’t reverse any of it. If anything, he’ll double down harder just to stop the Greens from outflanking them on the crazy scale. This is a contest between two factions of extremists fighting over who gets to finish the job of wrecking the country. And as always, the British people are the ones being thrown off the cliff.
0:10

Starmer resigning changes nothing. The mass migration, the rape gangs, the migrant crime, the green energy scam, these are Labour Party policy. Not just “Keir things.” So his replacement won’t reverse any of it. If anything, he’ll double down harder just to stop the Greens from outflanking them on the crazy scale. This is a contest between two factions of extremists fighting over who gets to finish the job of wrecking the country. And as always, the British people are the ones being thrown off the cliff.

Rothmus 🏴

25,648 次观看 • 2 个月前

Dynamic workflows are a generalization of harnesses, automations, loops, routing, and graphs. It's the most powerful feature I have built into my agent orchestrator. Supports all kinds of patterns that leverage different agent backends (claude, codex, pi, hermes,...). It's a meta-harness approach that unlocks new forms of test-time compute. Example of use cases it supports: > LLM councils to get different perspectives from LLMs or plan more intensively > Dynamically routing tasks to different agents based on needs (e.g., cost efficiency and optimal intelligence) > Advisor/Judge + executor workflows and pretty much any complex graph-based pattern required by the task. I find it especially useful for long-running work and code reviewing. > Agent teams that talk to each other if needed for the task. I like to use this for AI editing, artifact creation, and other creative tasks. And I am sure it supports so many things that I haven't discovered yet. I got inspired by the dynamic workflow feature released by the Claude Code team. I had actually built it earlier this year but wanted to generalize it across different agent backends. I think this is going to become more popular in the coming days. I will share more of my findings soon.
0:11

Dynamic workflows are a generalization of harnesses, automations, loops, routing, and graphs. It's the most powerful feature I have built into my agent orchestrator. Supports all kinds of patterns that leverage different agent backends (claude, codex, pi, hermes,...). It's a meta-harness approach that unlocks new forms of test-time compute. Example of use cases it supports: > LLM councils to get different perspectives from LLMs or plan more intensively > Dynamically routing tasks to different agents based on needs (e.g., cost efficiency and optimal intelligence) > Advisor/Judge + executor workflows and pretty much any complex graph-based pattern required by the task. I find it especially useful for long-running work and code reviewing. > Agent teams that talk to each other if needed for the task. I like to use this for AI editing, artifact creation, and other creative tasks. And I am sure it supports so many things that I haven't discovered yet. I got inspired by the dynamic workflow feature released by the Claude Code team. I had actually built it earlier this year but wanted to generalize it across different agent backends. I think this is going to become more popular in the coming days. I will share more of my findings soon.

elvis

32,623 次观看 • 1 个月前

It's nail night! This little man has huge paws and fast growing nails. He's been carefully introduced to the Dremel and doesn't fear it all. However sitting still long enough is a challenge for him so instead of spending a lot of time getting them really short, he gets just a little buzzed off and a bite of chicken or other high value treat every nail. Because it's a fast session they get done at least once a week. We start off with a quick sniff and paw touches on the Dremel so he's reminded it's a safe thing and essential to getting the treats. As he gets older and has more patience, he'll get longer, better trims and they won't need to be done as often but short sessions now keep him from getting frustrated and starting to really dislike the process. You really don't want an adult 120 pound rottweiler hating nail trims. #PuppyLife 🎥 Kenji doing his pre-pawdicure routine
0:14

It's nail night! This little man has huge paws and fast growing nails. He's been carefully introduced to the Dremel and doesn't fear it all. However sitting still long enough is a challenge for him so instead of spending a lot of time getting them really short, he gets just a little buzzed off and a bite of chicken or other high value treat every nail. Because it's a fast session they get done at least once a week. We start off with a quick sniff and paw touches on the Dremel so he's reminded it's a safe thing and essential to getting the treats. As he gets older and has more patience, he'll get longer, better trims and they won't need to be done as often but short sessions now keep him from getting frustrated and starting to really dislike the process. You really don't want an adult 120 pound rottweiler hating nail trims. #PuppyLife 🎥 Kenji doing his pre-pawdicure routine

Team Servicerottie🇨🇦🐕‍🦺🦽

18,772 次观看 • 7 个月前