วีดีโอ

Does it have the mobile application

Output tokens? Will this give short responses like every other llm?

Guess what the largest position in Michael Burry's (featured in The Big Short movie) portfolio is. Yes, it's Alibaba.

But I can't get 7b models to do well on 1k tokens lol

That's awesome!!!

love it, thanks

Thank you, excellent video. I am intensively using Qwen since a year and it gets better and better with every new release. And let us do not forget, that the DeepSeek R1 open-source models are also based on Qwen.

Learn how to build with LLMs, RAG, and AI Agents in my new courses here: dair-ai.thinkific.com/ Use code TH-cam20 to get an extra 20% off.

Thanks for the review. How is it compared to windsurf?

You could make it request quotes from different suppliers. Let's say you need a new car insurance or something?

Can it control other programs other than the web?

Ask him to trade forex

ask it to play a game of GeoGuessr

Please comment on what specific tasks you would like me to try. I will be doing another video on different use cases and I will be taking notes on you suggestions.

How far can you go while taking control? Can you just start something completely different like watching a youtube video? Effectively using it as a virtual machine?

I wish you would have shown it actually doing data entry. I want it to receive form data and then enter it into s spreadsheet :D Most of your use cases can be done faster just using regular GPT with web search capabilities.

Nice tool. What if a paper contains images/graphs or other unstructured format?

Learn how to build with LLMs, RAG, and AI Agents in my new courses here: dair-ai.thinkific.com/ Use code TH-cam20 to get an extra 20% off.

Learn how to build with AI in my new courses here: dair-ai.thinkific.com/ Use code TH-cam20 to get an extra 20% off. Limited-time offer!

❤

please make a video for tool/function calling agents with Deepseek-R1

I'll be watching this one very closely. Thanks for the video

Someone at some point is going to generate an executable and run it. I wonder what could go wrong.

Would you say it's better than o1 when it comes to your personal testing?

Hi Elvis, I'm in your online community (dair-ai).. loving your amazing job.. \o/

Hello! I have a question about Anthropic's prompt caching that I couldn't find in the documentation. Is it possible to share a cache between different prompts? For example, uploading a file in Prompt 1 and then accessing that same cached file in Prompt 2?

Hi please i am hoping an reply for my question. is it possible to fine tune a model using cloud service given and export the fine tuned model to deploy locally. If yes how can it be done

se la vie

Huge thanks please cover other techniques

Huge thanks

not support jetbrains

Great explanation with the multi-turn caching! Appreciate it keep up the great content🔥🤝

There is a 1/5 probability that by choosing a random answer, you will choose answer 3. Moreover, the only options that stand out after minimal filtering are 3 and 4 (the shortest or the longest candle), the rest are negligible, I believe that even ChatGPT wouldn't choose 1, 2, or 5. So, the real probability of finding the correct answer, even for the most primitive language model, I would say is somewhere around 50%. Out of 100 cases, in 50 of them, even the most primitive model would probably say it's 3. So, the answer is not impressive. A coin, if you flip it, would know to choose candle number 3 in 50% of the cases. But that doesn't mean the coin is good at math or that it's smart. So the fact that it answered with 3 is not at all impressive to me. What is impressive is the following logic, which seems to make sense and be correct.

So clear that it is easy to remember, thank you bro!

I am impressed by the quality of your content and how you properly provide context and data to ensure we can keep track of all the details provided.

I appreciate you taking the time to share your experience. However, this is not a tutorial. You must educate yourself on the pedagogy of creating tutorials before uploading additional content.

Great !

Awesome' Very useful content.

Have you try notebookllm plus? If so how is it ?

If you don't compare to the regular model then how do you know its working. I am yet to see a TH-cam video of a fine tuned model that works.. btw with the training loss greater than validation loss, it seems like over fitting and it wouldn't work

Can you create a custom Gem and share it with targeted people? I want an assistant that I only share with key people that I know. Any help would be greatly appreciated. Thank you!

I have had mixed results personally. I found o1-preview did not reliably produce better results. It frequently brushed over implementation details and was unwilling to get in to deep and complex topics.

I agree that results are inconsistent, but I'm puzzled by the supposition that it is cursor itself that is the cause of the problem you identified at the end with the model. Isn't that code being generated by the model on the back end and then sent back to cursor? So if it got the wrong model, wouldn't that be the LLMs fault, not cursor's? Or is cursor intercepting the token flow from the model and purposely injecting the wrong model? I would think that it's the model failing to follow the instructions consistently, rather than cursor intercepting and screwing the results. Not sure but that would be my first guess. I also agree that you really need to be a developer already to work effectively with these tools because they are inconsistent and will lead inexperienced developers down torturous paths if they don't already know programming well. For example, I asked cursor (sonnet) to create a somewhat complex t-sql routine. It created something extremely complicated because it didn't realize it could use a pre-existing t-sql function that comes shipped with SQL Server. But I knew. So I told it that its solution was way more complicated than necessary, and mentioned the function. It replied, "You're absolutely right!" and went ahead and reduced the 30 lines of code it had created to the 3 lines it actually needed. Etc, etc, etc, etc. But I don't blame cursor for this. It is the LLMs (sonnet, gpt4, etc) that cannot produce reliable results. Lastly, despite the flaws of the LLMs, because I am an experienced programmer who knows how to avoid the pitfalls presented by the LLMs, I am able to use the tools and save a tremendous amount of time. So they are good. It's just that you have to already be an expert to actually gain a real advantage from them. Newbies are likely to fall into every pit the models spit out at them, simply because they don't know better and will assume the model is doing things correctly. Bad assumption, sorry. Anyway, great video. I was wondering about agent. I think I'll give it a test drive. Thanks.

a pif ?

Coders digging their own grave