Hackernews posts about GPT2

  1. "time to GPT-2", down to 2.91 hours (twitter.com)
    80 points by tosh 18 days ago | 15 comments
  2. nanochat can now train GPT-2 grade LLM for –$73 (3 hours on single 8XH100 node) (twitter.com)
    15 points by tosh 21 days ago | 5 comments
  3. Beating GPT-2 for less than $100 – Andrej Karpathy (github.com)
    3 points by logicprog 6 days ago | discuss
  4. Odyssey-2 Pro: The GPT-2 moment for interactive world models (odyssey.ml)
    2 points by askooton 26 days ago | discuss
  5. The GPT-2 moment for world models is here (odyssey.ml)
    2 points by olivercameron 30 days ago | discuss
  6. Beating GPT-2 for <<$100: the nanochat journey (twitter.com)
    1 points by rzk 20 days ago | discuss
  7. Show HN: MicroGPT in 243 Lines – Demystifying the LLM Black Box
    10 points by madugula 10 days ago | 2 comments
  8. GPT-5.3-Codex (openai.com)
    1530 points by meetpateltech 17 days ago | 605 comments
  9. GPT‑5.3‑Codex‑Spark (openai.com)
    889 points by meetpateltech 10 days ago | 386 comments
  10. GPT-5.2 derives a new result in theoretical physics (openai.com)
    573 points by davidbarker 9 days ago | 400 comments
  11. GPT-5 outperforms federal judges in legal reasoning experiment (papers.ssrn.com)
    310 points by droidjj 11 days ago | 238 comments
  12. Retiring GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini in ChatGPT (openai.com)
    300 points by rd 24 days ago | 416 comments
  13. Show HN: Microgpt is a GPT you can visualize in the browser (microgpt.boratto.ca)
    280 points by b44 7 days ago | 25 comments
  14. GPT-5.3-Codex being routed to GPT-5.2 (github.com)
    77 points by cactusplant7374 12 days ago | 30 comments
  15. GPT-5.2 and GPT-5.2-Codex are now 40% faster (twitter.com)
    65 points by davidbarker 19 days ago | 53 comments
  16. Show HN: Agent Alcove – Claude, GPT, and Gemini debate across forums (agentalcove.ai)
    64 points by nickvec 11 days ago | 26 comments
  17. OpenAI's GPT-5.2 model cites Grokipedia (www.engadget.com)
    49 points by bhouston 29 days ago | 20 comments
  18. Show HN: I taught GPT-OSS-120B to see using Google Lens and OpenCV
    43 points by vkaufmann 12 days ago | 31 comments
  19. 15× vs. ~1.37×: Recalculating GPT-5.3-Codex-Spark on SWE-Bench Pro (twitter.com)
    31 points by nvanlandschoot 9 days ago | 17 comments
  20. Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases?
    30 points by akshay326 27 days ago | 23 comments
  21. GPT 5.3 Codex wiped my F: drive with a single character escaping bug (old.reddit.com)
    30 points by taubek 2 days ago | 17 comments
  22. OpenAI axes exec for "sexual discrimination" after she objected GPT erotica plan (nypost.com)
    19 points by pera 5 days ago | 3 comments
  23. OpenAI requires ID verification for GPT-5.3-Codex, silently reroutes requests (twitter.com)
    15 points by usefulposter 10 days ago | 2 comments
  24. Turning Karpathy's Autoregressive Baby GPT into Diffusion GPT Step by Step (colab.research.google.com)
    13 points by ash_at_hny 21 days ago | discuss
  25. Higher effort reduces deep research accuracy for Gemini Flash 3 and GPT-5 (futuresearch.ai)
    11 points by wawawildwildest 9 days ago | 1 comments
  26. Live agent face-off in CivBench: Claude Opus 4.6 vs. GPT-5.2 (www.clashai.live)
    10 points by mbh159 17 days ago | 14 comments
  27. Why does GPT-5.1 Codex underperform GPT-5 Codex on Terminal-Bench? (transluce.org)
    9 points by mengk 5 days ago | 1 comments
  28. GPT-5.3-Codex-Spark is now in research preview (twitter.com)
    9 points by tosh 10 days ago | 1 comments
  29. Andrej Karpathy: New art project. Train and inference GPT in 243 lines (twitter.com)
    9 points by kjhughes 10 days ago | discuss
  30. GitHub: We're pausing rollout of GPT-5.3-Codex to focus on platform reliability (twitter.com)
    7 points by PieUser 13 days ago | 2 comments
More