Uploaded November 2023 | Updated September 2026, 2 weeks ago
#gpt4 #promptengineering #compiler
I have built the most expensive CPU ever by compiling arbitrary C code to LLVM intermediate representation (LLVM-IR), then sending every single instruction to OpenAI's GPT API.
A true marvel of engineering.
Check out course: wandb.me/yannic-course
Repo: github.com/yk/llmvm
OUTLINE:
0:00 - Intro
0:55 - Sponsor Time
2:20 - FizzBuzz
4:05 - Computers 101
6:30 - Building a CPU on top of a large language model
7:40 - Compilers & LLVM
10:05 - From parsers to virtual machines
12:10 - GPTVM - A VM powered by GPT
15:55 - Snek!
18:40 - Going beyond AGI - Introducing Chad GPT VM
Links:
Homepage: ykilcher.com
Merch: ykilcher.com/merch
YouTube: youtube.com/c/yannickilcher
Twitter: twitter.com/ykilcher
Discord: ykilcher.com/discord
LinkedIn: linkedin.com/in/ykilcher
If you want to support me, the best thing to do is to share out the content :)
If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this):
SubscribeStar: subscribestar.com/yannickilcher
Patreon: patreon.com/yannickilcher
Bitcoin (BTC): bc1q49lsw3q325tr58ygf8sudx2dqfguclvngvy2cq
Ethereum (ETH): 0x7ad3513E3B8f66799f507Aa7874b1B0eBC7F85e2
Litecoin (LTC): LQW2TRyKYetVC8WjFkhpPhtpbDM4Vw7r9m
Monero (XMR): 4ACL8AGrEo5hAir8A9CeVrW8pEauWvnp1WnSDZxW7tziCDLhZAGsgzhRQABDnFy8yuM9fWJDviJPHKRjV4FWt19CJZN9D4n
#gpt4 #promptengineering #compiler
I have built the most expensive CPU ever by compiling arbitrary C code to LLVM intermediate representation (LLVM-IR), then sending every single instruction to OpenAI's GPT API.
A true marvel of engineering.
Check out course: wandb.me/yannic-course
Repo: github.com/yk/llmvm
OUTLINE:
0:00 - Intro
0:55 - Sponsor Time
2:20 - FizzBuzz
4:05 - Computers 101
6:30 - Building a CPU on top of a large language model
7:40 - Compilers & LLVM
10:05 - From parsers to virtual machines
12:10 - GPTVM - A VM powered by GPT
15:55 - Snek!
18:40 - Going beyond AGI - Introducing Chad GPT VM
Links:
Homepage: ykilcher.com
Merch: ykilcher.com/merch
YouTube: youtube.com/c/yannickilcher
Twitter: twitter.com/ykilcher
Discord: ykilcher.com/discord
LinkedIn: linkedin.com/in/ykilcher
If you want to support me, the best thing to do is to share out the content :)
If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this):
SubscribeStar: subscribestar.com/yannickilcher
Patreon: patreon.com/yannickilcher
Bitcoin (BTC): bc1q49lsw3q325tr58ygf8sudx2dqfguclvngvy2cq
Ethereum (ETH): 0x7ad3513E3B8f66799f507Aa7874b1B0eBC7F85e2
Litecoin (LTC): LQW2TRyKYetVC8WjFkhpPhtpbDM4Vw7r9m
Monero (XMR): 4ACL8AGrEo5hAir8A9CeVrW8pEauWvnp1WnSDZxW7tziCDLhZAGsgzhRQABDnFy8yuM9fWJDviJPHKRjV4FWt19CJZN9D4n
![[ML News] DeepMinds Flamingo Image-Text model | Locked-Image Tuning | Jurassic X & MRKL](https://i.ytimg.com/vi/smUHQndcmOY/mqdefault.jpg "[ML News] DeepMinds Flamingo Image-Text model | Locked-Image Tuning | Jurassic X & MRKL
#flamingo #mlnews #tech
Your updates directly from the state of the art in Machine Learning!
OUTLINE:
0:00 - Intro
0:30 - DeepMinds Flamingo: Unified Vision-Language Model
8:25 - LiT: Locked Image Tuning
10:20 - Jurassic X & MRKL Systems
15:05 - Helpful Things
22:40 - This AI does not exist
References:
DeepMinds Flamingo: Unified Vision-Language Model
https://www.deepmind.com/blog/tackling-multiple-tasks-with-a-single-visual-language-model
https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/tackling-multiple-tasks-with-a-single-visual-language-model/flamingo.pdf
https://twitter.com/Inoryy/status/1522621712382234624
LiT: Locked Image Tuning
https://ai.googleblog.com/2022/04/locked-image-tuning-adding-language.html
https://google-research.github.io/vision_transformer/lit/
Jurassic X & MRKL Systems
https://www.ai21.com/blog/jurassic-x-crossing-the-neuro-symbolic-chasm-with-the-mrkl-system#reading
https://arxiv.org/pdf/2205.00445.pdf
https://arxiv.org/pdf/2204.10019.pdf
https://studio.ai21.com/jurassic-x
StyleGAN Human
https://stylegan-human.github.io/
https://github.com/stylegan-human/StyleGAN-Human?utm_source=pocket_mylist
https://huggingface.co/spaces/hysts/StyleGAN-Human
Helpful Things
https://github.com/rish-16/grafog
https://huggingface.co/bertin-project/bertin-gpt-j-6B
https://github.com/pytorch/torchdistx
https://pytorch.org/torchdistx/latest/fake_tensor.html
https://github.com/Netflix/vectorflow?utm_source=pocket_mylist
https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/
https://twitter.com/DeepMind/status/1517146462571794433
https://github.com/ai-forever/mgpt
https://github.com/cleanlab/cleanlab
https://efficientdlbook.com/?utm_source=pocket_mylist
https://minihack-editor.github.io/
https://mugen-org.github.io/
https://www.amazon.science/blog/amazon-releases-51-language-dataset-for-language-understanding
https://github.com/phuselab/openFACS?utm_source=pocket_mylist
https://medium.com/pytorch/avalanche-and-end-to-end-library-for-continual-learning-based-on-pytorch-a99cf5661a0d
This AI does not exist
https://thisaidoesnotexist.com/
Links:
Merch: https://ykilcher.com/merch
TabNine Code Completion (Referral): http://bit.ly/tabnine-yannick
YouTube: https://www.youtube.com/c/yannickilcher
Twitter: https://twitter.com/ykilcher
Discord: https://ykilcher.com/discord
BitChute: https://www.bitchute.com/channel/yannic-kilcher
LinkedIn: https://www.linkedin.com/in/ykilcher
BiliBili: https://space.bilibili.com/2017636191
If you want to support me, the best thing to do is to share out the content :)
If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this):
SubscribeStar: https://www.subscribestar.com/yannickilcher
Patreon: https://www.patreon.com/yannickilcher
Bitcoin (BTC): bc1q49lsw3q325tr58ygf8sudx2dqfguclvngvy2cq
Ethereum (ETH): 0x7ad3513E3B8f66799f507Aa7874b1B0eBC7F85e2
Litecoin (LTC): LQW2TRyKYetVC8WjFkhpPhtpbDM4Vw7r9m
Monero (XMR): 4ACL8AGrEo5hAir8A9CeVrW8pEauWvnp1WnSDZxW7tziCDLhZAGsgzhRQABDnFy8yuM9fWJDviJPHKRjV4FWt19CJZN9D4n")
![[ML News] Chips, Robots, and Models](https://i.ytimg.com/vi/tRavLU8Ih4A/mqdefault.jpg "[ML News] Chips, Robots, and Models
OUTLINE:
0:00 - Intro
0:19 - Our next-generation Meta Training and Inference Accelerator
01:39 - ALOHA Unleashed
03:10 - Apple Inks $50M Deal with Shutterstock for AI Training Data
04:28 - OpenAI Researchers, Including Ally of Sutskever, Fired for Alleged Leaking
05:01 - Adobes Ethical Firefly AI was Trained on Midjourney Images
05:52 - Trudeau announces $2.4billion for AI-related investments
06:48 - RecurrentGemma: Moving Past Transformers for Efficient Open Language Models
07:15 - CodeGemma - an official Google release for code LLMs
07:24 - Mistral AI: Cheaper, Better, Faster, Stronger
08:08 - Vezora/Mistral-22B-v0.1
09:00 - WizardLM-2, next generation state-of-the-art-LLM
09:31 - Idefics2, the strongest Vision-Language-Model (VLM) below 10B!
10:14 - BlinkDL/rwkv-6-world
10:50 - Pile-T5: Trained T5 on the Pile
11:35 - Model Card for Zephyr 141B-A39B
12:42 - Parler TTS
13:11 - RHO-1: Not all tokens are what you need
14:59 - Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs
References:
https://twitter.com/ayzwah/status/1780263768968273923
https://ai.meta.com/blog/next-generation-meta-training-inference-accelerator-AI-MTIA/?utm_source=twitter
https://twitter.com/soumithchintala/status/1778087952964374854?t=Mb-mQvm4YIZ35pVpEijs6g&s=09
https://deepnewz.com/tech/apple-inks-50m-deal-shutterstock-ai-training-data
https://twitter.com/TolgaBilge_/status/1778598047821291793?t=zInlPDRZzozcz7-pjFSnyA&s=09
https://twitter.com/javilopen/status/1778821749792034911?t=oGLiMj6GQdKTuM6GbiYrAg&s=09
https://twitter.com/paulg/status/1781329523155357914?t=vCQT2mJf5BbtjdN1BMFYFQ&s=09
https://twitter.com/RichardSocher/status/1776706907295846628
https://www.cbc.ca/news/politics/federal-government-ai-investment-1.7166234
https://arxiv.org/pdf/2404.07839
https://huggingface.co/blog/codegemma
https://mistral.ai/news/mixtral-8x22b/
https://twitter.com/MistralAILabs/status/1780606904273702932?t=JlSCcYulpJL74pNJbtSZag&s=09
https://huggingface.co/Vezora/Mistral-22B-v0.1
https://huggingface.co/Vezora/Mistral-22B-v0.2
https://twitter.com/WizardLM_AI/status/1779899325868589372?t=l0Fd-4mfdtz3np_gALKaLA&s=09
https://twitter.com/_philschmid/status/1779922877589889400?t=7q1xg1LRy80mV8JGRm4aqA&s=09
https://huggingface.co/BlinkDL/rwkv-6-world
https://blog.eleuther.ai/pile-t5/
https://huggingface.co/HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1
https://huggingface.co/MaziyarPanahi/zephyr-orpo-141b-A35b-v0.1-GGUF
https://twitter.com/reach_vb/status/1778138382633140276?t=Mb-mQvm4YIZ35pVpEijs6g&s=09
https://arxiv.org/pdf/2404.07965
https://arxiv.org/pdf/2404.05719
https://sambanova.ai/blog/samba-coe-the-power-of-routing-ml-models-at-scale
https://www.microsoft.com/en-us/research/project/vasa-1/
https://twitter.com/twelve_labs/status/1780939765405065254?t=5ONxSzdwnghsKcwq3IPmEQ&s=09
https://drive.google.com/file/d/1Av5jpsbH3g09TRD1PfRh0nLsYrN_iu7_/view
https://arxiv.org/pdf/2404.12387
https://arxiv.org/abs/2404.12241
https://arxiv.org/pdf/2404.12241
https://twitter.com/Alon_Jacoby/status/1780650122382049596
https://audiodialogues.github.io/
https://os-world.github.io/
https://ai.meta.com/blog/openeqa-embodied-question-answering-robotics-ar-glasses/?utm_source=twitter&utm_medium=organic_social&utm_content=video&utm_campaign=dataset
https://arxiv.org/pdf/2404.07503
https://arxiv.org/pdf/2404.06654
https://twitter.com/amanrsanger/status/1779620682340704386?t=UnOronFwkESwAXiE0i0R4A&s=09
https://huggingface.co/datasets/xai-org/RealworldQA
https://github.com/PygmalionAI/aphrodite-engine
https://github.com/jina-ai/reader/?tab=readme-ov-file
https://r.jina.ai/https://x.com/elonmusk
https://r.jina.ai/https://github.com/jina-ai/reader
https://github.com/rogeriochaves/langstream
https://twitter.com/mvpatel2000/status/1777891913313440215?t=m5POrtTTS33tgwmRztQj3w&s=09
https://github.com/databricks/megablocks
https://github.com/nus-apr/auto-code-rover
https://github.com/nus-apr/auto-code-rover/blob/main/preprint.pdf
https://twitter.com/karpathy/status/1683143097604243456?t=7V_ApJFbjrm4TbxM5n3nXA&s=09
https://twitter.com/karpathy/status/1777427944971083809?t=s6xYQmYkhQyiFU65Fwq9tw&s=09
https://github.com/BasedHardware/Friend
https://twitter.com/argmaxinc/status/1781382688819282132?t=vCQT2mJf5BbtjdN1BMFYFQ&s=09
https://twitter.com/awnihannun/status/1778519566437794109?t=8N5PjwlKJpGotTx_HXZQrQ&s=09
https://twitter.com/Prince_Canuma/status/1776399292036501898
https://pytorch.org/blog/torchtune-fine-tune-llms/
If you want to support me, the best thing to do is to share out the content :)")
