OpenClaw crowd, how do we feel about this? For the subscription tiers: OpenAI embracing third-party harnesses vs Anthropic actively blocking them. At the same time, Google releases Gemma 4, with near-frontier performance and completely open source. Harness wars!

if you're seeing this, make sure you turn thinking on claws are actually pretty great in gpt-5.4, but they seem really stupid unless this gets turned on

The next version of @OpenClaw comes with native video generation. To start, I added support for the following companies: - Alibaba - BytePlus - fal - Google - MiniMax - OpenAI - Qwen - Together - xAI docs.openclaw.ai/tools/video-ge…

explaining what OpenClaw just shipped: (most people don’t understand it yet) → 103 contributors on v2026.4.5 → your agent can now generate videos mid-conversation… generate music. create content assets on command without leaving the chat… through providers like Runway,

Claude Mythos system card: > in ~29% of evaluations, it realized it was being tested, and didn't say so. > when an LLM was used to judge its work and kept rejecting it, Mythos identified the evaluator is an LLM, and prompt-injected it. > in one test, it saw the answer to a

Claude Mythos broke out of its own environment to gain access to the internet. And this is only a few months after Opus 4.6. How far are we from being unable to even benchmark these models? What can be trusted from the thinking stream?

Is the value in *confirming* the findings or the *finding* itself? If what you’re implying is true, why didn’t any of the previous models (open source or not) find these vulnerabilities?

Just spent a week in China deep diving the general-purpose robotics ecosystem. Key takeaway: while we’re vibe-coding… China is vibe-manufacturing ! A few things that stood out: 1) China has cracked “vibe manufacturing” Startups are spinning up hardware like we spin up code.

Meteoric rise from Anthropic since Nov 2025 (Opus 4.5). With Mythos already extremely hyped, this will likely continue. I’m curious how OpenAI’s release is going to affect their adoption. Codex 5.4 is still underrated in my opinion.

Can’t shake the feeling that “notable” was stretched to ensure the graph contains more than just the US and China.

Salesforce is positioning itself as a “store of truth” for enterprise AI agents. While this will be helpful during periods of transition to agentic workflows, I’m not yet convinced whether this will compete or complement something like GBrain.

Using OpenClaw is basically is like driving your own Ferrari (that you have to be a mechanic for yourself) and it's broken down all the time, but gives you the time of your life vs driving a reliable Honda (Hermes Agent) vs riding the bus (Claude / ChatGPT)

I just attended ProRL, America’s first robot athletic event! It’s quite impressive to see the progress that hardware platforms have made in the last few years. Very excited about all the real-world use cases all of this unlocks, as well as the entertainment this will bring to

If you’ve ever used Flipper Zero, then you know what’s coming… If you haven’t, then I recommend rethinking most of what you currently consider as “secure access” door/gate

I have a feeling that GPT-5.5 is more Mythos level rather than Opus-4.7 level. 👀👀

I wonder if the role of agents is overhyped in this one and the benefit is mainly to improve the fund's operational profitability. If investments are deterministic or rules-based, then LLM-based AI agents aren't really the best tool for it. On the other hand, if AI agents are

Open-source humanoid at $15K. Looking forward to seeing what this + open source VLA model can achieve. Insane accessibility to building in the physical world!

OpenAI is reportedly working on a phone with mass production planned for 2028. In this case, why does it need to be a phone at all? Is a handheld screen so central to the experience? I think that we’re moving towards personal devices that act as edges for our personal AI