Perhaps the best review of OAI dev-days I have seen in last few days! Btw how does it feel to keep saying "Prompt Engineering" in this dev-day, haha. In my opinion, OpenAI's recent dev day was impressive for its focus on practical developer tools rather than hype or new model releases. This approach wisely shifts attention from the irrational race of releasing marginally better models just to beat benchmarks. Instead, we should focus on building useful applications with our current, capable language models. I particularly appreciated the introduction of "evals," emphasizing real-world utility over abstract metrics. Standout features like model fine-tuning and structured output empower developers to create impactful solutions for end-users. This event seems to bring a welcome balance to the AI industry, prioritizing developer needs and practical applications. My hope is that this direction will ease the competitive pressure, allowing companies to focus on empowering developers who ultimately bring AI benefits to end-users - which is what truly matters.
I really dont see anything amazing out of OpenAI... They are doing what others are already doing, however they are same price, sometimes even more expensive.
![The moment we stopped understanding AI [AlexNet]](http://i.ytimg.com/vi/UZDiGooFs54/mqdefault.jpg)
The new llama 3.2 1B is great for the query rewrite in rag. It’s fast and does a pretty decent job.
Perhaps the best review of OAI dev-days I have seen in last few days! Btw how does it feel to keep saying "Prompt Engineering" in this dev-day, haha. In my opinion, OpenAI's recent dev day was impressive for its focus on practical developer tools rather than hype or new model releases. This approach wisely shifts attention from the irrational race of releasing marginally better models just to beat benchmarks. Instead, we should focus on building useful applications with our current, capable language models. I particularly appreciated the introduction of "evals," emphasizing real-world utility over abstract metrics. Standout features like model fine-tuning and structured output empower developers to create impactful solutions for end-users. This event seems to bring a welcome balance to the AI industry, prioritizing developer needs and practical applications. My hope is that this direction will ease the competitive pressure, allowing companies to focus on empowering developers who ultimately bring AI benefits to end-users - which is what truly matters.
Cool video, what screen recorder are you using btw
its screen.studio
I am waiting for you to develop advanced voice mode chrome extension for web chatgpt after this feature of API. Please do bro.
Pretty sure everyone is talking about.
I really dont see anything amazing out of OpenAI... They are doing what others are already doing, however they are same price, sometimes even more expensive.