I want to know if Deep Research can do my taxes for me..... if it can (for an LLCs taxes) and detail all of the places and timing of filing requirements and all that other stuff.... then it would be very worth it to me.
Great point about the environment, I feel like there're so many issues regarding configuration and incompatibility (especially with python), no way an AI can figure it out (without tools that could point it to documentation)
Guys come on. It's literally a matter of typing a second question to the model after the research all you have to do is enter your question. Or for that matter, just tell deep research to condense all of your research down into two paragraphs in the first place because my brain can't handle reading more than than three sentences. You don't have to read all of it, Are you serious right now? This is the first time you've used AI?
I could respond to this in the Discord where my opinions won’t date so badly 😊 but my in experience AI is not yet capable of building a full app with no ‘looking at code’ as it constantly forgets large sections of what it previously coded. Yes, it built that new functionality really well, and broke and simply removed 3 things it previously built. I still very much need to be, at the very least, a prompter and copy paste monkey (and in reality much more) Edit: Granted, thinking and listening further, a framework with decent modelling from a user could be the way to go, maybe my early 2000s training in UML wasn’t a waste after all 😅
For long output you can use xml tags with o1 (repo prompt) For claude you can use MCPs and ask it to modify snippets of your local code (instead of the whole file)
Great episode, but just one thing... At 14:09 with that spinning ball example, if you look closer the o3 model is simulating gravity physics whereas the R1 is not.
![WE GOT ACCESS TO GPT-3! [Epic Special Edition]](http://i.ytimg.com/vi/iccd86vOz3w/mqdefault.jpg)
![WE GOT ACCESS TO GPT-3! [Epic Special Edition]](/img/tr.png)
At the moment my 2 most used LLMs are o3-mini-high and Gemini 2.0 Flash Thinking - I am waiting for o3:)
When is 03 mini free?.
People in developing countries don't care if their data is sent to America or China.
Gemini so much better
Great episode, the song was amazing. It's also nice to have longer episodes ❤
Great show
I want to know if Deep Research can do my taxes for me..... if it can (for an LLCs taxes) and detail all of the places and timing of filing requirements and all that other stuff.... then it would be very worth it to me.
Great point about the environment, I feel like there're so many issues regarding configuration and incompatibility (especially with python), no way an AI can figure it out (without tools that could point it to documentation)
Guys come on. It's literally a matter of typing a second question to the model after the research all you have to do is enter your question. Or for that matter, just tell deep research to condense all of your research down into two paragraphs in the first place because my brain can't handle reading more than than three sentences. You don't have to read all of it, Are you serious right now? This is the first time you've used AI?
I could respond to this in the Discord where my opinions won’t date so badly 😊 but my in experience AI is not yet capable of building a full app with no ‘looking at code’ as it constantly forgets large sections of what it previously coded. Yes, it built that new functionality really well, and broke and simply removed 3 things it previously built.
I still very much need to be, at the very least, a prompter and copy paste monkey (and in reality much more)
Edit: Granted, thinking and listening further, a framework with decent modelling from a user could be the way to go, maybe my early 2000s training in UML wasn’t a waste after all 😅
I'm equally tired of OpenAI and Musk getting all the attention...
Please spare us from this onslaught.
For long output you can use xml tags with o1 (repo prompt)
For claude you can use MCPs and ask it to modify snippets of your local code (instead of the whole file)
Great episode, but just one thing... At 14:09 with that spinning ball example, if you look closer the o3 model is simulating gravity physics whereas the R1 is not.
Google pro 1206 and ow 0206 are my goto.not just cos they are free
you guys drastically underestimate how lazy people are -- when it comes to clicking a mouse and keyboard.
I was the 100th like
Yawn
Got researcher did this month's ago