Grok DeepSearch vs. ChatGPT DeepResearch

Thoughts? I’m a Lead Researcher and would love to hear about your experience using both of these tools. I love ChatGPT, but if Grok is better than I’m open to using both.

Thanks!

4 Likes

I asked ChatGPT and it swore it was better. :sweat_smile:

Seriously, though, I’d test them myself if I were you as it really comes down to what you’re using it for…

All that said, I’ve found ChatGPT DeepSearch helpful and gave up on Grok a while ago…

3 Likes

Hi!

I haven’t touched Grok—and I truly have no intention to—but DeepResearch is freaking awesome.

I tried it yesterday with two different prompts. The first took 10 minutes, the second took a whopping 20 minutes to complete! I actually thought there was an error with the latter, but was delighted to discover the research finished up even after I closed my laptop and changed locations.

Both essays are well reasoned, well cited, and well written. Both have insights and perspectives I hadn’t considered, and both are imminently usable for my purposes. Both seem “less hallucinatory,” and accurate.

I haven’t tried limiting or recommending sources for research, but the idea seems pretty straight forward.

It would be really neat if there were a way of limiting said sources to a for-pay service, for example high-quality subscription services like Reuters or the New York Times (where I, the user, would be responsible for maintaining a subscription for the premium sources).

3 Likes

Both are insufficient in my opinion.

I have found that both models use whatever is SEO’d, lacking discern for marketing material. This is before they become a commodity.

Keep in mind that a lot of informational pages survived on ad revenue. With AI this dimension is eliminated. The only public information pages available in the future will be privately funded for the sole purpose of influencing AI.

What would be more important (in my opinion) is to be able to collaborate with a model in selecting the sources, and then performing a “deep research”, not just letting a model choose for itself

3 Likes

I haven’t used Grok, but I used DeepResearch this morning for a research question involving 17th century Calvinist Protestant theology and its resistance to Quakers in New Netherland. This hatred of Quakers was also shared by the Puritans of New England, but I’m researching the Dutch. I was impressed with the depth of resources it dug up, and the number of references. It was a thorough analysis, citing details from multiple sources, allowing me to draw my own conclusions and summarize the results. Bottom line if anyone is interested: Quakerism was truly subversive in the 17th century, granting women equality, refusing to pay tithes to the church, and embracing pacifism, which was in direct opposition to both Protestantism and Catholicism, which were actively engaged in domination, subjugation, warfare and enslavement. It was viewed with the same kind of revulsion and horror by the powerful elites of the time, the same way DEI is viewed today by the elites, as a subversive force threatening their hold on power. And money.

3 Likes

Hi
It was excited experiment, and i significantly contributed in developing both of them and they gave me certificates related to my achievements

1 Like

Brilliant thank you, that makes a lot of sense.

both get a lot of stuff wrong in my opinon

grok atlesat 10x better. i started using it a week ago, and was blown away by it. first, its answers are just… simply better. chatgpt keeps halucianting and forgetting crucial details in my prompts. grok doesnt… grok also has no input character limit as far as i know, and i have inputted over 50k or something before. also, the deep"ER" search feature is great. it analyzed over 104 sorces at once, and gave me extremely accurate stock predictions somehow.

For me my customized GPT is better. 4.5 and Grok3 told me so:). If you’d like to test it with a question, I ‘ll share the output. I have to work on it more before opening it up, but its definately different.