been reading about openai in 2022 and it’s hilarious how desensitised they were to the tech before launching chatgpt, expecting only a few thousand users.
they thought it would just be a “lowkey research preview” lol
training a gpt-2 model (124m) was a fun learning experience.
i think this week i'm going to implement the llama architecture and see how the updates to the transformer affect the training run.
the uk media have been mostly silent on the public uprising, but you can feel them spiralling now that elon’s blasting to the world that remigration is the only solution.
several years late to the party, but attempting my first pretraining run of a small gpt-2 on 8x a100s.
not like it will be useful but has been a fun learning experience.
ok so now that i'm back expect posts on:
- ml/ai i probably don't understand
- failing at other areas of software
- project updates i won’t finish
- endless uk doomposting
definitely worth sticking around.
ok i didn't expect work to get so chaotic and me to get so lazy while at the end of this shitty project.
will ship this week no excuses.
gonna go back to some ml stuff after, learning web dev is draining me rn.
spent like 2 hours using various llms to attempt to fix a bug before giving up and deciding to read the actual docs.
fixed it myself within 10 mins.
just read the docs guys.
decided to self-host my project instead of using something like vercel.
feels right to understand the infrastructure if i'm already this deep into learning web dev.
first obstacle: confronting my docker skill issue.
six months into this job and only today noticed chatgpt has been blocked the entire time for "data security".
meanwhile i've been pumping company data into claude and gemini daily without issue. they really think they solved the problem.
back on the frontend journey after my break. building a simple rag app to learn svelte.
python backend and testing pydantic ai to handle llm functionality.
hoping to ship in the next week or two.
been off the tl for a couple weeks. life got busy and work had me travelling around the country.
i don't think i'm built for daily posting.
we're back now though, got a couple projects in the pipeline.
71 Followers 906 FollowingWhite House Tech Support,Passionate about accelerating the world's transition to sustainable energy,& ensuring humanity's long-term survival and prosperity⚡️🚀
447 Followers 2K FollowingI like rockets and tech/AI. I shitpost sometimes. Musician (jack of all instruments, master of none). Studying Computer Science.
24K Followers 10K FollowingFormer Quant Investor, now building @lumeraprotocol
(formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
325K Followers 3K FollowingNVIDIA Director of Robotics & Distinguished Scientist. Co-Lead of GEAR lab. Solving Physical AGI, one motor at a time. Stanford Ph.D. OpenAI's 1st intern.
30K Followers 123 FollowingMechanistic Interpretability lead DeepMind. Formerly @AnthropicAI, independent. In this to reduce AI X-risk. Neural networks can be understood, let's go do it!
20K Followers 9K FollowingProgramme Director @ARIA_research | accelerate mathematical modelling with AI and categorical systems theory » build safe transformative AI » cancel heat death
6K Followers 740 FollowingPreserving episodic memory and logic through ontology shifts. Integrating cyborg layers. Learning the place of knowing in the all. Disjuncting on unknowns.
263K Followers 666 FollowingBuilding with AI agents @dair_ai • Prev: Meta AI, Galactica LLM, Elastic, PaperswithCode, PhD • I share insights on how to build with AI Agents ↓
5K Followers 843 FollowingEvery age, it seems, is tainted by the greed of men. Rubbish to one such as I, devoid of all worldly wants. — I work on HPC and making AI run faster.
56K Followers 853 FollowingFiguring out AI @allen_ai, open models, RLHF, fine-tuning, etc
Contact via email.
Writes @interconnectsai
Wrote The RLHF Book
Mountain runner
22K Followers 52 FollowingCommunity account for sharing ClaudeCode related projects and releases. Views/shares independent from @AnthropicAI positions.