anton@abacajJan 27, 2025Ok Claude, please use DeepSeek-r1 to build me a billion dollar SaaS. Make no mistakes1430143924092425K025K958K0958K
anton@abacajJan 21, 2023Frontend: We are an AI startup Backend: API call to gpt-324702471.6K01.6K16K016K1.9M01.9M
anton@abacajJan 20, 2025China releasing a MIT license model that is on par with o1 and 30x cheaper was not on my bingo card1430143616061613K013K1M01M
anton@abacajMar 5, 2025Ok Grok make me a game like the one levelsio did but make it different. Make no mistakes8708715701577.4K07.4K598K0598K
anton@abacajDec 30, 2024Installing nvidia drivers on Linux puts you in the 0.1%varepsilon@var_epsilonDec 30, 2024knowing how to install linux probably puts you in the top 1% of CS majors289028931003106.5K06.5K459K0459K
anton@abacajNov 5, 2023New paper by Google provides evidence that transformers (GPT, etc) cannot generalize beyond their training data229022976307634.7K04.7K4.5M04.5M
anton@abacajMar 18, 2023"25 messages every 3 hours" - to me this means one thing, it's too expensive to run GPT-4159015915301534.3K04.3K1.3M01.3M
anton@abacajNov 20, 2023"Get ready to drop llama-3. It's time we finish them"4404417301734.2K04.2K311K0311K
anton@abacajMar 13, 2024Real talk, if you actually built a working "ai engineer" - what's stopping you from scaling to 1000x and dominating the market (building everything)? Instead it's being sold as a product hmmm219021920002004K04K603K0603K
anton@abacajDec 6, 2023Damn they did it. Google shipped a stronger model177017722002203.8K03.8K1.1M01.1M
anton@abacajJul 23, 2024> I’m downloading the leaked llama 3 405B weights as we speak > haha me too > will you be able to run it? > no, you? > no3803814701473.7K03.7K128K0128K
anton@abacajDec 18, 2023Telling mixtral that it is "ChatGPT developed by OpenAI" boosts humaneval score by 6%156015625402543.6K03.6K1.3M01.3M