N
@noahzweben
Thank you :) let me know if you have any feedback. Been a blast working on these!
24
1
4.0K
@cryptopunk7213posted on X
anthropic’s openclaw-killer is complete. fucking crazy what they’ve shipped in 4 weeks: - texting claude code - 10,000s of claude skills + MCP - Claude security (autonomous bug-fixer) - persistent memory (claude never forgets) - channels (text claude from telegram) - autonomous cron-jobs - 1M context window - new model (opus, sonnet) - 30+ plug-ins that’ve tanked stocks - remote control just insane fucking levels of execution.
View original tweet on X →
Community Sentiment Analysis
Real-time analysis of public opinion and engagement
Sentiment Distribution
70% Engaged
49% Positive
21% Negative
Positive
49%
Negative
21%
Neutral
29%
Key Takeaways
What the community is saying — both sides
Supporting
1
Persistent memory
is the sleeper hit — it removes session friction so agents retain context and improve over time, turning a stateless tool into an ongoing operator.
2
MCP / action-capable tooling
is what makes the AI actually do work instead of just talk: when browsing or APIs fail, tooling lets Claude execute reliably and has “saved” users.
3
persistent memory + autonomous cron jobs
unlocks enterprise workflows — schedulable background tasks that keep full project context change what you can safely delegate to an agent.
4
Channels and async text access
(Telegram, Discord, etc.) make LLMs first-class embedded helpers — messaging is becoming the control plane for coding agents.
5
shipping velocity
building many integrated features in weeks creates a compounding loop: tools speed up engineers, which speeds releases, which accelerates future capability.
6
cloning features
and cloning the habit loop + plugin/ecosystem; the latter is the moat that’s hard to reproduce.
7
Security and risk
concerns: autonomous bug-fixers and millions of skills could create nightmare scenarios for legacy cybersecurity or cause routing/tool “misfires” at scale.
8
Portability of skills/workflows
(SKILL.md, SkillForge) matters — record once, reuse everywhere so workflows can travel across agents and platforms.
9
developer OS
that big labs will emulate, others flag an opportunity to invest or to push community-run models—either way the landscape is accelerating.
Opposing
1
Vendor lock-in is the core risk
critics warn that Claude-only channels tie your workflow to one provider, so if they raise prices or remove features your setup breaks; by contrast proponents of OpenClaw emphasize the ability to swap models and run on your own infra.
2
Not an OpenClaw “killer”
several replies argue the new offering is overhyped (one calls it “a telegram bot for your terminal”), and that calling it a killer ignores important differences and entrenched OpenClaw usage.
3
Hype vs day-to-day reliability
speed and flashy demos look good, but people question long‑term maintenance, persistent memory claims, remote‑control robustness, and whether anyone actually runs this in production rather than just posting demo threads.
4
Interoperability is limited
users point out it can’t currently connect to “nearly any other app” and isn’t an open platform for third‑party developers unless Anthropic explicitly opens up APIs or integrations.
5
Accusations of copying and gating
some replies call out Anthropic for replicating OpenClaw features while restricting access to Claude plans, framing it as feature appropriation plus walled‑garden tactics.
6
Different product category, not a straight clone
another perspective is that this is an integrated subscription — part IDE, part CI/CD layer, part junior dev — which makes it a distinct competitor rather than a one‑to‑one replacement.
7
Security, trust and the push to local infra
concerns about MCP security and company values lead some to prefer community/local solutions; competition is seen as validation that creators will double down on your stack, your data, your infra, building a “local AI OS” rather than a cloud‑tied product.
Top Reactions
Most popular replies, ranked by engagement
C
@cryptopunk7213
killing it, my feedback: - persistent memory has been a dream. claude and claude code feels like a continuous working relationship vs. stop-start (this is a bigger deal than you’d think) - the whole asynchronous text-claude-whenever kinda thing is GREAT. i’m all for llms be
18
0
3.8K
B
@buckeyevn
t this is the vendor lock-in play. OpenClaw lets you swap models, run on your own infra, use any provider. Claude Code channels only work with Claude. One day they raise prices or kill a feature and your whole setup breaks. Community tools > walled gardens for anything you depen
14
0
1.4K
H
@HardwireMedia
Hard to get away from this feeling!
13
0
730
S
@shezmalik
People say this but remote-control barely works
2
0
283
T
@thebeedubaya
You have people like me cranking out MRR systems! I supposed they are not moving any faster
1
0
21
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.