Permalink

1987 - Backpropagation
1989 - CNN
1998 - LeNet
2007 - CUDA
2009 - ImageNet
2011 - ReLU
2012 - AlexNet
2014 - GANs
2015 - TensorFlow
2016 - PyTorch
2017 - Transformers
2018 - GPT
2019 - GPT-2
2020 - GPT-3
2022 - InstructGPT
2022 - ChatGPT
2023 - GPT-4, Claude, Llama
2024 - Gemini

Post image
Permalink

AI Factory -> Pre-trained -> Transformers -> Finetune -> Mistral -> reinforcement learning ->  Mistral_Instruct

Permalink

LLM : Compress public human knowledge into public domain matrices

 ᛡ 2026 Transformers World

Post image
Permalink

when you train an LLM with local GPUs

Post image
Permalink
Post image
Permalink

old day finetune mistral

Post image
Post image
Permalink

Your favorite terminal with native git worktrees.

ghostree.dev

Post image
Post image
Permalink

Cloudflare Sandbox

Sandboxes are ideal for building AI agents that need to execute code, interactive development environments, data analysis platforms, CI/CD systems, and any application that needs secure code execution at the edge. Each sandbox runs in its own isolated container with a full Linux environment, providing strong security boundaries while maintaining performance.

npm i @cloudflare/sandbox

sandbox.cloudflare.com

Post image
Post image
Permalink
Build, cable, and automate. Assemble racks, servers, and switches, connect enough capacity for each app, earn money from processed data, and unlock gear and bigger customers with XP and reputation. Follow colored data packets as your center comes to life.

store.steampowered.com/app/4170200/Data_Center

Post image
Post image
Post image
Permalink

Anemoiapolis exists in the uncanny valley of American architecture -- endless carpeted floors and lonely malls.
Escape for your life, and enjoy the scenery.

youtube.com/watch?v=ZzmENbOqfaE

Permalink

liminal dream

Post image
Post image
Post image
Permalink
Permalink

Docker Sandboxes

MicroVM-Based Isolation for Coding Agents

- Each agent runs inside a dedicated microVM
- Only your project workspace is mounted into the sandbox
- Hypervisor-based isolation significantly reduces host risk
- Agents can install system packages, run services, and modify files
- Workflows run unattended, without constant permission approvals
- Coding agents can build and run Docker containers inside the MicroVM
- They have no access to the host Docker daemon
- If an agent goes off the rails, delete the sandbox and spin up a fresh one in seconds

docker.com/blog/docker-sandboxes-run-claude-code-and-other-coding-agents-unsupervised-but-safely

Post image
Permalink

Deno Sandbox

You don’t want to run untrusted code (generated by your LLMs, your users LLMs, or even hand written by users) directly on your server. It will compromise your system, steal your API keys, and call out to evil.com. You need isolation.

Deno Sandbox gives you lightweight Linux microVMs (running in the Deno Deploy cloud) to run untrusted code with defense-in-depth security. You create or programmatically via our JavaScript or Python SDKs, and they boot in under a second. You can also interact with them via SSH, HTTP, or even open a VS Code window directly into the sandbox.

deno.com/blog/introducing-deno-sandbox

Post image
Permalink

Vibe

Vibe is a quick, zero-configuration way to spin up a Linux virtual machine on Mac to sandbox LLM agents

LLM agents are more fun to use with --yolo, since they're not always interrupting you to approve their commands.

Sandboxing the agent in a VM lets it install/remove whatever tools its lil' transformer heart desires, without wrecking your actual machine.

Virtualization is more secure against malicious escapes than containers or the MacOS sandbox framework.

Containers on MacOS require spinning up a virtual machine anyway.

The binary is < 1 MB.

The entire implementation is in one ~1200 line Rust file.

github.com/lynaghk/vibe

Post image
Permalink
Permalink
Post image
Close
Fullscreen image