HN Explorer

Ask HN: Examples of O1 pro DeepResearch proving its worth?

5 points by ChaitanyaSai ↗ HN
Hey all! We use all the LLMs fairly heavily. We see their promise and are aware of their pitfalls. Wondering if DeepResearch really is a major step up compared to the rest or even our jerry-rigged prompt chains. If you've used it and have found a significant upside, can you please share your experience. An actual output would be wonderful. Thank you!

4 comments

[ 4.4 ms ] story [ 20.4 ms ] thread
psyklic
Stratechery just published detailed specific examples of when it's useful and not: https://stratechery.com/2025/deep-research-and-knowledge-val...
thundergolfer
> my definition of Artificial Super Intelligence (ASI) is the ability to come up with the tasks in the first place).

It’s remarkable how the business and strategy types keep adding new definitions of these things. “Coming up with new tasks” isn’t superintelligence, that’s just general intelligence. Ben Thompson even calls out that OpenAi annoyingly has like a dozen AGI definitions directly after. Yet he still adds to the pile and the confusion.

The G means “as intelligent as a typical human”. The S means “significantly more intelligent than a typical human”.

Any system which doesn’t realize you can’t wash your hands if you are missing both arms is not earning that G.