Anthropic just released 5 workshops on building self-improving agentic systems from scratch: 00:00 - Ship your first Claude agent 36:44 - Self-improving agents (tools,skills) 1:21:25 - Build memory for Claude agents 1:49:47 - Set up a proactive agent 2:11:31 - Make your agent autonomous These 3-hours of Claude workshops... will replace 20 paid agentic courses. Watch today, then read article below on how to build a self-improving agentic system with Fable 5.show more

combining skills and memory for agents is where most people get stuck, cool to see both covered here

Yeah, this is the ultimate guide to building agentic systems, bro. Definitely worth turning into a book.

Self-improvement is a fantasy turned into a lose-lose situation for users. You don’t need to rewrite your skills on every cycle like tokens are nothing unless you have a very good reason to. Anthropic is literally losing thousands of dollars in compute on negative unit economics.

3 hours to replace 20 courses is crazy thats like one costco hotdog replacing your entire grocery budget

Great breakdown. Workshops like these are incredible for individual agents, but the real paradigm shift happens when you orchestrate them under a single, unified operating system. ​That's exactly why we are building Eleven. The ecosystem is shifting. ⚡

the self-improving agents workshop is the one worth your time. memory is still the hardest part nobody's actually solved well

bro dropped a whole bootcamp for free and people still out here paying $200 for udemy courses 💀

3 hours of pure alpha from the people who built Claude, completely free. definitely worth checking out

and if I run the first agent on Fable 5, will it be expensive?

The agent system was complicated for me, but I figured it out

Is this available in YouTube?

This is a solid resource. Anthropic has been putting out some really useful learning content lately.

skip the memory one at 1:21 and youll wonder why your agent forgets everything mid task

Codez feels sharp Nicely framed

This should make things much easier

Thanks for the timecodes

This is the part most people miss. Agents need boring workflows, memory, and feedback loops, not one magic prompt.

Those timestamps make it seem like a marathon. Hope there’s enough depth for real insights.

This is the kind of content that's worth bookmarking.

Is this in YouTube?

this is basically an agent builder roadmap in three hours

Good video and great article

Watching the workshop isn't the edge. Shipping the agent before the video ends is.

This is the part I think most people skip. Self-improving agents need a place to write failures back. Otherwise the agent improves for one session, then Monday morning starts from vibes again.

Self-improving agents make memory the critical layer. If an agent can update its tools, skills and behavior, then we need to know: what changed why it changed which source supported it whether other agents can verify it Autonomy without verifiable memory is fragile. This is where trusted context layers like OriginTrail / ethereum:0xaa7a9ca87d3694b5755f213b5d04094b8d0f0a6f become interesting. @0xCodez @origin_trail @TraceLabsHQ @CryptoRyan17

Self improving agents from scratch is exactly what builders need right now. Starting today.

That's amazing

official Anthropic workshops on this stack outperform most paid courses because the framing matches how the model was actually trained. worth the 3 hours before paying anyone

Hey there vibecoders, check out our new agentic coding IDE’s 🚀. It’s 800% faster than others. Follow and Drop us a DM for a free pro trial account.

anthropic is making a serious effort to make agentic systems accessible to more people

That’s the real leverage. The companies that learn faster than everyone else, because their products continuously improve from user feedback, will be incredibly difficult to catch.

@caspaofficial calling for backup

This is absolutely amazing

Awesome, may I know where can I find out these courses by Claude? I would like to learn more.

@xdownloaderbot download this