Hackernews posts about GPT3

GPT3 is a revolutionary AI model that uses machine learning to generate human-like text based on input prompts, allowing for applications such as language translation, content creation, and natural language processing.

1. Scaling Laws for LLMs: From GPT-3 to o3 (cameronrwolfe.substack.com)
   1 points by UCdallasGA 17 days ago | discuss
2. Fuzzy API composition: querying NBA stats with GPT-3 and Statmuse and LangChain (www.geoffreylitt.com)
   1 points by sebg 30 days ago | discuss
3. Show HN: OmiAI – A highly opinionated AI SDK with auto-model selection (github.com)
   1 points by yoeven 25 days ago | discuss
4. Revenge of the GPT Wrappers: Defensibility in a world of commoditized AI models (andrewchen.substack.com)
   135 points by kiyanwang 14 days ago | 44 comments
5. OpenAI Roadmap Update for GPT-4.5 and GPT-5 (twitter.com)
   58 points by gigaArpit 9 days ago | 13 comments
6. OpenAI's GPT-5 Bait and Switch (www.vincentschmalbach.com)
   38 points by vincent_s 8 days ago | 20 comments
7. GPT-4o-powered cleaning robot (built in 4 days) (twitter.com)
   27 points by Jannik-TRLC 26 days ago | 10 comments
8. AI Models Like GPT-4o Change Without Warning. Here's What You Can Do About It (www.libretto.ai)
   13 points by aickin 2 days ago | 5 comments
9. GPT4 level intelligence fell 1000x in 18 months (twitter.com)
   11 points by stevekrouse 2 days ago | 14 comments
10. Claude is More Anxious than GPT (www.lesswrong.com)
    10 points by higuidebot 11 days ago | 13 comments
11. GPT-5 Could Be Coming as Soon as May (www.theverge.com)
    6 points by dev-jayson about 21 hours ago | 1 comments
12. Tülu 3 405B – GPT-4o performance with open-source data, code, and recipes (twitter.com)
    6 points by rvnx 22 days ago | discuss
13. A prompt-trained DeepSeek R1 70B can perform better than GPT-o1 using AdalFlow (colab.research.google.com)
    4 points by meame2010 16 days ago | 4 comments
14. Jailbroken: Finding 50,000 Legal Hallucinations in GPT-4o with RL (generalanalysis.com)
    4 points by rhavaei 22 days ago | 2 comments
15. Want to Train Your Own GPT-Style Model? – Step-by-Step Notebook (github.com)
    4 points by kevinpdev 6 days ago | 1 comments
16. ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheaper (twitter.com)
    4 points by moondistance 29 days ago | discuss
17. GPT-4 writes better edit summaries than human Wikipedians (en.wikipedia.org)
    3 points by tonymet 11 days ago | discuss
18. Gptel: Mindblowing integration between Emacs and ChatGPT (www.blogbyben.com)
    3 points by sebg 23 days ago | discuss
19. GPT from Scratch with MLX (www.rayfernando.ai)
    3 points by skidrow 29 days ago | discuss
20. I scraped 1.6M jobs with GPT4o-mini (hiring.cafe)
    2 points by hamed_n 29 days ago | 2 comments
21. Filtering spam with GPT4o-mini for $0.00008 per email (blog.abctaylor.com)
    2 points by arcza 2 days ago | 1 comments
22. DeepSeek: A 100-Person Team Built an AI Model Rivaling GPT-4 (chinacompany.substack.com)
    2 points by zacfire 4 days ago | 1 comments
23. GPT 4 IaC usecases IRL (www.goncharov.xyz)
    2 points by ultral 9 days ago | 1 comments
24. Distilled DeepSeek R1 Outperforms Llama 3 and GPT-4o in Classifying Error Logs (rootly.com)
    2 points by sylvainkalache 2 days ago | discuss
25. New GPT-4o Copilot code completion model available now (github.blog)
    2 points by e2e4 3 days ago | discuss
26. Sam Altman on X: "trying GPT-4.5 has been much more of a "feel the AGI" moment…" (twitter.com)
    2 points by enpera 4 days ago | discuss
27. Converting GPT-4o into a reasoning model (wandb.ai)
    2 points by byyoung3 5 days ago | discuss
28. Show HN: Quiz Me GPT: AI generated trivia questions for any topic you choose (wd13.ca)
    2 points by willd13 10 days ago | discuss
29. Revenge OF THE GPT WRAPPERS: Defensibility in a world of commoditized AI models (twitter.com)
    2 points by leonagano 17 days ago | discuss
30. Show HN: Who is hiring jobs using GPT-4/embedding models for relevant jobs (registry.jsonresume.org)
    2 points by thomasfromcdnjs 20 days ago | discuss