Cofounder @ Firstloop, an ai-native software development agency. We're hiring, DMs open

Denver, CO
Joined September 2013
Jake Colling retweeted
Here you go, link in the next post
Tibo
@thsottiauxSep 10
Replying to @tldraw @sama
Show me. Post tomorrow by EOD
104
80
52
2,089
386,673
Jake Colling retweeted
working on this model, with this team, with all of the infrastructure we built to get here, is the most rewarding thing i've ever been a part of still can't believe i get to wake up every day and work on this stuff with these people and put all the code for free on the internet and tweet about it for all of you guys the @primeintellect mission, vision, and master plan is just getting started :)
Replying to @PrimeIntellect
INTELLECT-3 is a 106B parameter Mixture-of-Experts model trained with both SFT and RL on top of the GLM 4.5 Air Base model. Both stages, including multiple ablations, were carried out on a 512-GPU H200 cluster over the course of two months.
58
93
7
1,196
116,161
Jake Colling retweeted
We're all in on coding agents. Sometimes we have five working at the same time. Read about our high-level strategy here 👇 firstloop.ai/blog/coordinati…
1
2
438
Jake Colling retweeted
If anything, this trade war has shown has clueless you average American is about where their food and goods come from and how they get to their house.
19
7
1
125
17,923
Jake Colling retweeted
I can't believe what i just found. an unremovable swing with Locktite™️ nuts and 3/8" galvanized chain. crazy
this is the most insane arc. SF parks and rec is doing the opposite of their logo
24
8
9
249
32,437
Replying to @JacobColling
I'm in 💪🏻👍🏻
2
1
1
178
What are you all using for docs / articles in @nextjs ? Trying to roll by own with .mdx is working, but really finnicky and much more work for features than I want to be putting in. Struggling the most with building navigation and routing from file/folder structure.
4
1
4
1,192
Introducing FastMLX 🎉 Thrilled to unveil FastMLX, a high-performance server designed for hosting MLX models on your Mac. FastMLX supports both Vision Language Models (VLMs) and Language Models (LMs) 🚀 Key Features: • OpenAI-compatible API • Asynchronous Calls • Dynamic Model Loading: Load MLX models on-the-fly or use pre-loaded models for better performance. • Image Processing Capabilities: Handle both text and image inputs for versatile model interactions. Get started now: > pip install -U fastmlx Please leave us a star and send a PR :) github.com/Blaizzy/fastmlx
11
26
6
172
29,007
Using Athena to accelerate PE Deal Rooms 👇
3
4
450
Jake Colling retweeted
Happy to share Browserbase with the world today. We help AI applications browse the web. And we just raised $6.5 million to do it. Now, we're opening signups to developers everywhere. I can't wait to see what you 🅱️uild.
122
62
60
593
305,175
Is this thing on?
1
1
1
3
914
Jake Colling retweeted
The mind-map returns... but for LLMs... Testing out Caret by @JacobColling .
1
1
12
547
A showcase of four Olympus interfaces in CoPilot Mode, with a toy example, saving and analyzing an OpenAI attendee list for an upcoming event. This video highlights the benefit of an ecosystem of copilot user interfaces. Depending on the work to be done, we can quickly pivot from... 1. Browsing the web with Athena -- with takeover control to get past a Luma login screen or solve a CAPTCHA. 2. Saving web content (parsed text and raw HTML) or full-page screenshots to the Olympus Drive. 3. Add this list to a sheet and download it to CSV or Excel. 4. Breaking the glass and accessing the raw contents in a Jupyter Notebook for advanced analysis from the Olympus Workbench.
1
1
5
912
Caret is my new favorite way of interacting with LLMs! It's an obsidian plugin that makes it easy to have branching conversations! I'm using it with @ollama with Llama3:8b and it works really well!
Big News: I’m going to open source Caret! Why, what that means for people who paid for early access, what’s next for Caret and what's next for me:
6
5
6
98
15,471
Jake Colling retweeted
Try Caret: an Obsidian plugin enabling local-first, privacy-preserving use of LLMs within your vault. With it's AI Canvas, customizable workflows, and chat integration, Caret transforms how you capture and evolve ideas, ensuring your data stays secure. caretplugin.ai/
Big News: I’m going to open source Caret! Why, what that means for people who paid for early access, what’s next for Caret and what's next for me:
2
7
1,973
Living the life with #Caret (poke @JacobColling). GPT-4o still has a lot to learn about le parler marseillais, though (poke @MedericGC)
1
1
6
877
Excited to announce today that I have acquired @homerow_app from @dexterleng It lets you navigate your entire Mac & the Web without a mouse. The best way to understand it is watch this demo/launch video I just cooked up.
39
12
4
154
30,938
Jake Colling retweeted
1/3 The Caret from @JacobColling for @obsdmd is a mind-blowing ! You can chat with anyone from a Markdown (MD) file in Obsidian using AI and visual tools. It's super fast! I can't wait to use the full potential of this tool!
1
1
3
551
Jake Colling retweeted
I did the same thing bought two license love what your doing
1
1
2
332
Jake Colling retweeted
Replying to @JacobColling
i love caret, gonna be buying another license 😁
1
1
5
653