Articles by anjneymidha
81

State of AI: An Empirical 100T Token Study with OpenRouter (openrouter.ai)

2

Flux2 VAE Research Report (bfl.ai)

1

How AI Works – A Primer (publish.obsidian.md)

2

Show HN: Komposer, AI image editor where the LLM writes the prompts (komposer.xyz)

117

UN report finds UN reports are not widely read (reuters.com)

2

Track and visualize LLM model performance over time (github.com/jacobphillips99)

1

Pufferlib: Simplifying reinforcement learning for complex game environments (github.com/pufferai)

35

America officially goes to war with Iran (twitter.com/america)

2

PSA: Codex crossed 350K GitHub PRs merged (twitter.com/anjneymidha)

42

[dupe] First Boeing 787 Dreamliner Crash (nytimes.com)

1

Mistral Compute (mistral.ai)

3

FLUX.1 Kontext:In-Context Image Generation and Editing in Latent Space [pdf] (sanity.io)

1

Inside The First Stargate AI Data Center (bloomberg.com)

1

Who Wins Nobel Prizes (construction-physics.com)

13

Verification, the Key to AI (2001) (incompleteideas.net)

20

OpenAI Announces New SWE-Lancer Benchmark (openai.com)

1

The SWE-Lancer Benchmark (twitter.com/openai)

6

Grok3 is first model to surpass 1400 on the Chat Arena benchmark (twitter.com/lmarena_ai)

1

How to Backdoor Large Language Models (sshh.io)

4

Kimi K1.5: Scaling Reinforcement Learning with LLMs (arxiv.org)

2

Flux runs 2x faster on 10GB VRAM with Nvidia FP4 (blackforestlabs.ai)

2

GPU Glossary (modal.com)

4

Ask HN: What projects are you hacking on this week for fun?

38

Ask HN: Stanford CS 153 help

1

I, Pencil [pdf] (fee.org)

2

The Brittle Lesson: Claude and Computer Use (anjneymidha.com)

2

Sharing new research, models, and datasets from Meta FAIR (meta.com)

13

Meta Spirit LM: Open multimodal language model that freely mixes text and speech (twitter.com/aiatmeta)

2

FLUX1.1 [pro] - New SOTA image model + public API from Black Forest Labs (twitter.com/bfl_ml)

1

A Preliminary Report on DisTrO [pdf] (github.com/nousresearch)

5

OpenAI Microscope (microscope.openai.com)

1

Indian Voters Are Being Bombarded with Deepfakes. Candidates Approve (wired.com)

2

Mistral CEO confirms new 8x22B model is Apache 2.0 (twitter.com/arthurmensch)

2

Discord Embedded App SDK is now live (discord.com)

2

Discord opens up to games and apps embedded in its chat app (theverge.com)

43

Chess-GPT's Internal World Model (github.com/adamkarvonen)

4

Amazon to invest upto $4B in Anthropic (twitter.com/anthropicai)

3

The Bitter Lesson (2019) (incompleteideas.net)

189

Llama2.c: Inference llama 2 in one file of pure C (github.com/karpathy)

1

Nintendo Family Office (y-n10.com)

1

Optimising computer systems with more generalised AI tools (deepmind.com)

257

Deepmind Alphadev: Faster sorting algorithms discovered using deep RL (nature.com)