
Brayden Levangie
@blevlabs
Followers
2K
Following
4K
Media
74
Statuses
2K
building cognitive superintelligence @LevangieLabs
San Francisco, CA
Joined August 2021
Real progress in AI will come in the form of decentralized, ambient superintelligence. It is the only way forward to have a true global impact.
OpenAI has been "rerouting" prompts to its old model 4o to a new "safety model." I've been reading a lot of posts from people who are pissed. They are ones who have dove head first into AI, building relationships with it that are, well, weird even to me. And I'm weird that way.
2
3
17
Great video from Prompt Engineering on YT regarding Workflows vs Agents Good clarity for anyone learning about agents right now, especially to not be confused by @OpenAI’s recent releases https://t.co/bONTYCrJIE
0
2
10
For our free newsletter this week, we ask whether autonomous systems should have a say in product roadmaps or policies. @IrenaCronin and I write this newsletter every week. Even when I am on vacation, like I am now. Hello from Barstow, on way to Los Angeles. Autonomous
9
15
92
Hiring for a Cognitive Agent Motivational Speaker role internally. We already have agents building their own prompting + tools, so this is all that's left.
1
0
4
@Scobleizer this is a good example of the labs building on the expectations the world has set for LLMs than trying to build real agentic intelligence. Goes back to the paradigm of chat ≠ actions I shared and the consequences of what that has done to how people innovate in AI.
0
1
4
If your agents are based on workflows (n8n) it is a sign you are building in the wrong direction. Disappointing, but expected esp. with the direction @OpenAI has went.
BREAKING 🚨: OpenAI is planning to announce Agent Builder on DevDay. Agent builder will let users build their agentic workflows, connect MCPs, ChatKit widgets and other tools. This is one of the smoothest Agent builder canvases I've used so far. The year of Agents 🤖
1
2
8
Well @Scobleizer looks like the finally released something. Interested to check it out.
Introducing Tinker: a flexible API for fine-tuning language models. Write training loops in Python on your laptop; we'll run them on distributed GPUs. Private beta starts today. We can't wait to see what researchers and developers build with cutting-edge open models!
0
0
4
At @DeveloperCamp_ we democratize innovation by putting the tools of enterprise R&D in everyone’s hands. Thanks to the efforts of our Community Sponsor @Deepinvent, the organization Cada is now our grant-writing partner. They help innovative companies secure paths to
13
31
80
We built a robot brain that nothing can stop. Shattered limbs? Jammed motors? If the bot can move, the Brain will move it— even if it’s an entirely new robot body. Meet the omni-bodied Skild Brain:
515
920
7K
Oh, never mind. That would have been the most useful part.
@jasteinerman Very cool, but getting access to the display would change everything. Hope that’s coming soon
1
0
2
Love this. Was going to try and get access to @theworldlabs, but I think I’ll go with the open-source option. Open-source research ftw!
Can we use video diffusion to generate 3D scenes? 𝐖𝐨𝐫𝐥𝐝𝐄𝐱𝐩𝐥𝐨𝐫𝐞𝐫 (#SIGGRAPHAsia25) creates fully-navigable scenes via autoregressive video generation. Text input -> 3DGS scene output & interactive rendering! 🌍 https://t.co/HBdrmU4Oqq 📽️ https://t.co/AQr0p4uWBZ
0
0
3
Thanks to @dom for organizing this incredible event and also huge thanks to @Scobleizer for inspiring me to get more into the SF community!
1
0
12
Thank you for the opportunity for such an amazing event! So excited I was able to help so many of the amazing developers at this hackathon and excited to be the grand prize winner!
Thrilled to announce our grand prize winner and our collaboration with @helpdotfun on their groundbreaking launch. We started with 9 decoys and 1 real token. The fees from the decoys fueled liquidity in the real token, igniting a massive surge in trading volume. In just a few
4
1
18
Do what I do: Make the agent have all the fundamental tools it needs to interact with an OS, have it write its own toolsets, define its behavior, and then create a better version of itself for you. Let the agents worry about their own tools. They are the ones using them!
New on the Anthropic Engineering blog: writing effective tools for LLM agents. AI agents are only as powerful as the tools we give them. So how do we make those tools more effective? We share our best tips for developers:
0
1
8
em-dash detected. sign they have not reached a breakthrough in llms yet. Joking aside, interesting paper. Hope to see more than a blog post soon!
Today Thinking Machines Lab is launching our research blog, Connectionism. Our first blog post is “Defeating Nondeterminism in LLM Inference” We believe that science is better when shared. Connectionism will cover topics as varied as our research is: from kernel numerics to
0
0
3