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。

30 hours. One agent run. Sonnet 4.5 handled a full-day coding job end-to-end with lower latency and fewer handoffs.

Superhuman AI

2,789 subscribers

124,727 次观看 • 1 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

2 条评论

Sihab Safin 的头像

Sihab Safin1 年前

Wait this is actually so cool ! I’ve been following your posts lately and they always make me stop and think. how do you even keep up with all these crazy AI updates every day? 👀

Superhuman AI 的头像

Superhuman AI1 年前

Appreciate that! 🙌 We try to stay ahead of the curve so you don’t have to. For even cooler updates, you can subscribe to our newsletter — it’s where we share the good stuff first 👀

相关视频

ICYMI, GitHub Copilot coding agent now works with Claude Sonnet 4.5 - which you can leverage through Visual Studio Code's deep coding agent integration.

ICYMI, GitHub Copilot coding agent now works with Claude Sonnet 4.5 - which you can leverage through Visual Studio Code's deep coding agent integration.

Visual Studio Code

33,996 次观看 • 1 年前

Claude Sonnet 4.5 vs 4.0 (measured on coding tasks): App from scratch → 4.5: ~20min, functional build. → 4.0: 40+min, lower quality. Simple code change → 4.5: 3 tool calls, correct edit. → 4.0: longer chain, same result. Complex refactor → 4.5: faster + tests passed. → 4.0: slower + failed output. 4.5 reduces wasted cycles—less time waiting, fewer retries, more working code.

Claude Sonnet 4.5 vs 4.0 (measured on coding tasks): App from scratch → 4.5: ~20min, functional build. → 4.0: 40+min, lower quality. Simple code change → 4.5: 3 tool calls, correct edit. → 4.0: longer chain, same result. Complex refactor → 4.5: faster + tests passed. → 4.0: slower + failed output. 4.5 reduces wasted cycles—less time waiting, fewer retries, more working code.

Augment Code

67,588 次观看 • 1 年前

This week metrics and logs join traces in Mastra Studio. • Logs: every event, searchable with full context • Traces: debug any run, end to end • Metrics: cost, latency, errors, and scores at a glance With all three pillars of observability in one place, you can now see the whole picture.

This week metrics and logs join traces in Mastra Studio. • Logs: every event, searchable with full context • Traces: debug any run, end to end • Metrics: cost, latency, errors, and scores at a glance With all three pillars of observability in one place, you can now see the whole picture.

🏴󠁧󠁢󠁥󠁮󠁧󠁿 Paul

10,509 次观看 • 7 个月前

🚨 One Prompt To A Fully Functional SaaS App And Complete Software Opus 4.7 -> front-end GPT 5.5 -> long running complex backend Gemini 3.5 - > AI chatbot embedded in the app Sonnet 4.6 -> scheduled task mask management Kimi 2.6 -> simple cron jobs One prompt will build AND test everything - complete end -to-end execution multi-agent systems where each agent specializes in a particular task - design, testing, engineering

🚨 One Prompt To A Fully Functional SaaS App And Complete Software Opus 4.7 -> front-end GPT 5.5 -> long running complex backend Gemini 3.5 - > AI chatbot embedded in the app Sonnet 4.6 -> scheduled task mask management Kimi 2.6 -> simple cron jobs One prompt will build AND test everything - complete end -to-end execution multi-agent systems where each agent specializes in a particular task - design, testing, engineering

Abacus.AI

1,571,927 次观看 • 5 个月前

"more than twice the speed" is underselling Haiku tbh built a way to directly compare Sonnet v Haiku 4.5 and it's roughly 3.5x faster, but the UX feels SO much better because Haiku stays inside the "flow window". obviously end to end latency varies a lot so Ant can't report a real number without production usage but you should try heads up comparisons

"more than twice the speed" is underselling Haiku tbh built a way to directly compare Sonnet v Haiku 4.5 and it's roughly 3.5x faster, but the UX feels SO much better because Haiku stays inside the "flow window". obviously end to end latency varies a lot so Ant can't report a real number without production usage but you should try heads up comparisons

swyx

171,675 次观看 • 1 年前

Sanji shares his Higgsfield Supercomputer playbook to run marketing end-to-end. He lets the agent analyze his niche, find the winning formula, generate 100+ variations, and run A/B tests at scale, with the agent learning from every run. How do you use Supercomputer for scaling?

Sanji shares his Higgsfield Supercomputer playbook to run marketing end-to-end. He lets the agent analyze his niche, find the winning formula, generate 100+ variations, and run A/B tests at scale, with the agent learning from every run. How do you use Supercomputer for scaling?

Higgsfield AI 🧩

115,336 次观看 • 5 个月前

Claude Sonnet 4.5 is now the default model for Augment Code. We’re rolling it out to all customers over the next 24 hours, where it will be available alongside Sonnet 4 (for a limited time) and GPT-5 in the model picker. Here’s how Sonnet 4.5 compares to Sonnet 4:

Claude Sonnet 4.5 is now the default model for Augment Code. We’re rolling it out to all customers over the next 24 hours, where it will be available alongside Sonnet 4 (for a limited time) and GPT-5 in the model picker. Here’s how Sonnet 4.5 compares to Sonnet 4:

Augment Code

29,679 次观看 • 1 年前

🚨BREAKING: Claude Sonnet 4.5 just dropped and it's INSANE It can literally work for 30 hours straight BY ITSELF If you aren't taking advantage of everything new here you're falling behind Here's everything you need to know about Sonnet 4.5 and how to get the most out of it:

🚨BREAKING: Claude Sonnet 4.5 just dropped and it's INSANE It can literally work for 30 hours straight BY ITSELF If you aren't taking advantage of everything new here you're falling behind Here's everything you need to know about Sonnet 4.5 and how to get the most out of it:

Alex Finn

93,249 次观看 • 1 年前

good morning /v1/chat/completions This is a test we ran overnight on TensorRT-LLM with modified kernels serving a custom 1B parameter model we trained for a customer ~200ms end-to-end latency (not TTFB, full request). Beats their current Cerebras stack on latency and quality

good morning /v1/chat/completions This is a test we ran overnight on TensorRT-LLM with modified kernels serving a custom 1B parameter model we trained for a customer ~200ms end-to-end latency (not TTFB, full request). Beats their current Cerebras stack on latency and quality

Sam Hogan 🇺🇸

43,655 次观看 • 9 个月前

Built from scratch by Grok 4.5 + Grok Build in UE5.8: a cyberpunk L-corner street with neon facades, rain, signs, and crowds walking through the scene. End-to-end, the run took 10.75M tokens, 36.5 minutes, and only ~$12.4 at API pricing. What impressed me most is not just the final render, but the process: the coding agent built it in many small steps, saved 30+ map checkpoints, verified the scene, and iteratively fixed issues autonomously. Congrats to the team for building such an impressive coding model and harness. This feels like a real step toward agentic 3D world creation. Elon Musk Andrew Milich skcd Yun-Ta Tsai #UnrealEngine

Built from scratch by Grok 4.5 + Grok Build in UE5.8: a cyberpunk L-corner street with neon facades, rain, signs, and crowds walking through the scene. End-to-end, the run took 10.75M tokens, 36.5 minutes, and only ~$12.4 at API pricing. What impressed me most is not just the final render, but the process: the coding agent built it in many small steps, saved 30+ map checkpoints, verified the scene, and iteratively fixed issues autonomously. Congrats to the team for building such an impressive coding model and harness. This feels like a real step toward agentic 3D world creation. Elon Musk Andrew Milich skcd Yun-Ta Tsai #UnrealEngine

SimWorld

645,980 次观看 • 3 个月前

Here’s how to build a Slack clone with Claude 3.7 Sonnet and Cursor Agent in v0.46. Perfect for *anyone* wanting to pick up AI coding tools. Learn how to build an app 100% from scratch with the newest AI coding workflows. Watch for 1hr+ of agentic coding.

Here’s how to build a Slack clone with Claude 3.7 Sonnet and Cursor Agent in v0.46. Perfect for anyone wanting to pick up AI coding tools. Learn how to build an app 100% from scratch with the newest AI coding workflows. Watch for 1hr+ of agentic coding.

Mckay Wrigley

212,638 次观看 • 1 年前

Real-time video agents are here. Today, we’re sharing how we built Runway Characters, allowing you to turn one image into a fully expressive, conversational video agent streaming at 24 frames per second in HD. With just 1.75 seconds of end-to-end latency. Learn more below.

Real-time video agents are here. Today, we’re sharing how we built Runway Characters, allowing you to turn one image into a fully expressive, conversational video agent streaming at 24 frames per second in HD. With just 1.75 seconds of end-to-end latency. Learn more below.

Runway

149,287 次观看 • 6 个月前

We are living in the most insane timeline. I just asked Claude Code (with Claude Sonnet 4.5) to develop an MCP Server (end-to-end) that allows me to programatically create n8n workflows from within Claude Code itself. Took about 10 mins!

We are living in the most insane timeline. I just asked Claude Code (with Claude Sonnet 4.5) to develop an MCP Server (end-to-end) that allows me to programatically create n8n workflows from within Claude Code itself. Took about 10 mins!

elvis

221,878 次观看 • 1 年前

Claude Cowork is cooked? With Claude Cowork, you’re still prompting one task at a time. LobeHub users deploy agent teams that run the entire workflow end-to-end.. 1.5x faster, 4x cheaper than Manus Tutorial below:

Claude Cowork is cooked? With Claude Cowork, you’re still prompting one task at a time. LobeHub users deploy agent teams that run the entire workflow end-to-end.. 1.5x faster, 4x cheaper than Manus Tutorial below:

Robin Delta

19,733 次观看 • 9 个月前

This is Claude Sonnet 4.6: our most capable Sonnet model yet. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also features a 1M token context window in beta.

This is Claude Sonnet 4.6: our most capable Sonnet model yet. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design. It also features a 1M token context window in beta.

Claude

7,612,085 次观看 • 8 个月前

📣 Anthropic's Claude Sonnet 5 is now generally available and rolling out in GitHub Copilot. Early testing for Claude Sonnet 5 showed: • strong results across a range of coding scenarios, particularly for performance on CLI-style tasks. • excellent prompt-cache utilization and competitive latency at lower effort levels. Try it out in the GitHub Copilot app ⬇️

📣 Anthropic's Claude Sonnet 5 is now generally available and rolling out in GitHub Copilot. Early testing for Claude Sonnet 5 showed: • strong results across a range of coding scenarios, particularly for performance on CLI-style tasks. • excellent prompt-cache utilization and competitive latency at lower effort levels. Try it out in the GitHub Copilot app ⬇️

GitHub

146,620 次观看 • 4 个月前

Cursor can now control your browser. Agent can take screenshots, improve UI, and debug client issues. Try our early preview with Sonnet 4.5.

Cursor can now control your browser. Agent can take screenshots, improve UI, and debug client issues. Try our early preview with Sonnet 4.5.

Cursor

1,549,377 次观看 • 1 年前

🚨 CANCEL YOUR SaaS Subscriptions - Build With AI Use top models including - Opus 4.7 & Sonnet 4.6 - GPT 5.5 high - Gemini 3.1 Pro Build entire apps, backends, APIs and mobile apps all using ENGLISH - The agent also has a built in Q/A engineer BUILD END TO END APPS WITH AI

🚨 CANCEL YOUR SaaS Subscriptions - Build With AI Use top models including - Opus 4.7 & Sonnet 4.6 - GPT 5.5 high - Gemini 3.1 Pro Build entire apps, backends, APIs and mobile apps all using ENGLISH - The agent also has a built in Q/A engineer BUILD END TO END APPS WITH AI

Abacus.AI

3,299,141 次观看 • 5 个月前

Just took my Tesla Cybertruck out with FSD Supervised v13.2.9, and it’s next-level! 🤖 The end-to-end driving network is so smooth – nailed lane changes and handled a tight two-lane road like a pro. The 36 Hz AI4 video inputs and low latency make it feel like the truck’s reading the road in real-time.

Just took my Tesla Cybertruck out with FSD Supervised v13.2.9, and it’s next-level! 🤖 The end-to-end driving network is so smooth – nailed lane changes and handled a tight two-lane road like a pro. The 36 Hz AI4 video inputs and low latency make it feel like the truck’s reading the road in real-time.

Tesla Owners Silicon Valley

66,852 次观看 • 1 年前