@felixrieseberg amazing. ability to select projects / mounted folders from Dispatch soon?
English
Alex Webbs
3.9K posts
Alex Webbs
@OnChainAlex
growth @Sei_FND
Boston Tham gia Kasım 2015
4.3K Đang theo dõi3.3K Người theo dõi
Claude Dispatch, a way for you to talk to Claude Cowork & Claude Code on your computer, is now available on all Teams plans.
It's off by default - if you don't see it in your sidebar, ask your team admin to turn it on.
English
Every time Anthropic launches a new Claude feature, and they do it like three times a week, I run to check the documentation. And every single time they have documentation ready. It is actually amazing. Really amazing.
English
Took the /remote-control magic and powered a single-long running session for Cowork Dispatch.
Coolest abilities:
1. Send files from local machine so you can work on PPTs on the go
2. Spawn sub-sessions on Desktop that you can drill down on
3. Chat about any local cowork session
Felix Rieseberg@felixrieseberg
We're shipping a new feature in Claude Cowork as a research preview that I'm excited about: Dispatch! One persistent conversation with Claude that runs on your computer. Message it from your phone. Come back to finished work. To try it out, download Claude Desktop, then pair your phone.
English
@trq212 amazing! can we get fast mode and ultrathink on code desktop too?
English
@zealoushacker we need a new memory & context management cookbook for cowork!🙏
English
@alexalbert__ is this limited to claude code or does it apply to cowork/desktop generally as well?
English
Tons of improvements shipped with this one:
- Opus 4.6 1M is now the default Opus model for Claude Code users on Max, Team, and Enterprise plans.
- No more long context price increase in the API.
- No beta header required in the API.
- Include up to 600 images in one request.
Claude@claudeai
1 million context window: Now generally available for Claude Opus 4.6 and Claude Sonnet 4.6.
English
Introducing Base44 Superagents.
AI agents built with managed infrastructure, secured by default, one-click integrations, and 24/7 execution from the start.
Everything is taken care of so you can focus on what your agent does, not how to get it running.
That means no API keys to juggle, no config files, no security setup, and no maintenance. We handle all of it.
Your Superagent connects to all the tools you already use in one click, runs on schedules and triggers, remembers context across sessions, acts proactively on your behalf, and keeps working around the clock.
All from wherever you already are, WhatsApp, Telegram, Slack, or your browser.
The AI agent everyone's been waiting for, with everything you need already built in.
We're excited to get this into your hands, so we're giving free credits to everyone who comments and reposts in the next 24 hours.
English
Humans don't use APIs, humans use UIs. APIs were built for programs to talk to programs. MCP is a standard so every AI doesn't have to independently figure out auth, capability discovery, and service little quirks for every single mf integration.
Protocols exist for interoperability not because individual endpoints are too hard.
English
Thank god MCP is dead
Just as useless of an idea as LLMs.txt was
It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs
Morgan@morganlinton
The cofounder and CTO of Perplexity, @denisyarats just said internally at Perplexity they’re moving away from MCPs and instead using APIs and CLIs 👀
English
'fix the implementation' doesn't go viral the way 'MCP is dead' does. So here we are.
Rhys@RhysSullivan
MCP sucking is a harness problem, not an MCP problem MCP unlocks behavior that is fundamentally impossible to get via CLI or APIs Bad auth, too much context usage, all get solved with an execution layer - your agent writes code to progressively discover and call tools
English
MCP sucking is a harness problem, not an MCP problem
MCP unlocks behavior that is fundamentally impossible to get via CLI or APIs
Bad auth, too much context usage, all get solved with an execution layer - your agent writes code to progressively discover and call tools
Garry Tan@garrytan
MCP sucks honestly It eats too much context window and you have to toggle it on and off and the auth sucks I got sick of Claude in Chrome via MCP and vibe coded a CLI wrapper for Playwright tonight in 30 minutes only for my team to tell me Vercel already did it lmao But it worked 100x better and was like 100LOC as a CLI
English
All of the complaints here about MCPs are valid but this is an unnecessarily binary framing.
The agent still needs to know what integrations exist and when to invoke them. Whether it's an MCP tool description or a filename in your saved library, you're still paying a discovery cost somewhere.
What you've actually built is a persistent environment with an Al agent that can write and run code. That's a cool product, just say that.
The honest framing is 'MCPs work well for the 80% of OOTB use cases & our product is great for the 20% that need custom logic and a persistent place to run it.'
That is a perfectly compelling product story, but instead we get 'MCP is fundamentally broken' which is just a product launch dessed as an architecture post. We need more honest framing not less!!
David@dzhng
English
Ya the auth is buggy, the context cost sucks and toggling is super annoying.
But MCP is still a new design space and has improved a lot even just over the last couple of months.
The auth story will continue to improve, context management will get smarter and the protocol will evolve.
All of the mcp is dead takes are lame and unproductive for being so conclusive for something so nascent. Ppl are talking their books.
English
MCP sucks honestly
It eats too much context window and you have to toggle it on and off and the auth sucks
I got sick of Claude in Chrome via MCP and vibe coded a CLI wrapper for Playwright tonight in 30 minutes only for my team to tell me Vercel already did it lmao
But it worked 100x better and was like 100LOC as a CLI
Morgan@morganlinton
The cofounder and CTO of Perplexity, @denisyarats just said internally at Perplexity they’re moving away from MCPs and instead using APIs and CLIs 👀
English
WE'RE NOT TALKING ENOUGH ABOUT HOW QUICKLY WE'RE GOING TO TOKENIZE EVERY FINANCIAL ASSET - Larry Fink
>BlackRock, Hamilton Lane, Apollo & more are onchain
>Enterprise giants like PayPal are issuing stablecoins
>US States are building their own stablecoin programs
>Toku is bringing the $50T payroll market onchain
>NYSE, ICE, DTCC are exploring blockchain rails
>AI agents are adopting x402 for payments
You can send a dollar globally in 400ms. You can send a dollar anywhere for $0.0005.
It has never been faster, cheaper, or easier to pay anyone, for anything, anywhere.
It's here. Global finance is onchain. And it moves faster on Sei.
English
We just raised $500M at an 11B valuation 🎉
To celebrate, we’re giving away 1,000 free credits so you can test our platform.
For the next 6 hours, comment “11B” below and we’ll DM you the credits (must follow) 👇
ElevenLabs@ElevenLabs
We raised $500M at an $11B valuation to transform how people interact with technology.
English
@ianlapham run multiple iterations simultaneously across tabs and toggle through refining each while waiting on the others to update, repeat.
when one version emerges as the best you fork that one a few times and switch to iterating on those.
English
the issue with purely vibe coding is that there is no flow state, by definition you need to multitask because you are always waiting interrupted between micro tasks
English
Tokenizing T-Bills was only a systems integration test to prove the rails work. Now comes the hard part: financializing the messy, illiquid Long Tail of RWAs.
Sei Development Foundation@Sei_FND
$33B in tokenized assets isn’t the product. It’s the proof of concept. So far, RWAs have largely solved putting already-liquid markets onchain. The real transformation is unlocking the assets that can’t trade at all. New opinion piece from @ChadHugghins and @OnChainAlex.
English
The future is concurrent. Speed creates chaos; engineering controls it.
New research addresses the unique dynamics of MCP.
Duplicate-Aware Payouts & Priority-DAGs break the monopoly and neutralize same-tick extraction.
Split the tips. Order by priority. Simple, robust, & fast.
Jay ($/acc)@jayendra_jog
Sei Labs just published the first ever paper on MEV in next-generation blockchains. Future blockchains, like Sei Giga, will use Multiple Concurrent Proposers. However, MCP introduces new types of MEV like tx stealing, proposer to proposer auctions and timing races.
English
非常荣幸加入 @Sei_FND 担任华语区负责人，未来将负责推动 SEI 华语生态建设，并希望凝聚更多对 SEI 感兴趣的伙伴，共同发展，期待得到大家的认可和支持。
关于我为什么选择 Sei，Sei 是少数以高吞吐量为核心战略的公链之一，致力于从底层重构打造 Web2 级响应能力的 L1。即将推出的 Giga 升级将并行 EVM 性能推向新高度。目标超 20 万 TPS、执行效率提升 50 倍、亚秒级最终确认，满足机构级金融应用需求。定位和底层技术在一众公链里面，具有非常高的发展潜力。
目前base香港，欢迎DM交流！
I’m joining the Sei Development Foundation as GM, Greater China. Fired up by Sei’s momentum across real-world finance and the upcoming Giga upgrade (targeting 200K+ TPS, and sub‑second finality) that unlocks real RWAs and payments use cases at scale. If we’re connected and you’re building in the region, let’s catch up and swap notes. DMs open. 🚀
中文