Comment by tptacek | Hacker News Reader

Comment by tptacek | Hacker News Reader

I did a trial run with Deep Research this weekend to do a comparative analysis of the comp packages for Village Managers in suburbs around Chicagoland (it's election season, our VM's comp had become an issue).

I have a decent idea of where to look to find comp information for a given municipality. But there are a lot of Chicagoland suburbs and tracking documents down for all of them would have been a chore.

Deep Research was valuable. But it only did about 60% of the work (which, of course, it presented as if it was 100%). It found interesting sources I was unaware of, and assembled lots of easy-to-get public data that would have been annoying for me to collect that made spot-checking easier (for instance, basic stuff like the name of every suburban Village Manager). But I still had to spot check everything myself.

The premise of this post seems to be that material errors in Deep Research results negate the value of the product. I can't speak to how OpenAI is selling this; if the claim is "subscribe to Deep Research and it will generate reliable research reports for you", well, obviously, no. But as with most AI things, if you get paste the hype, it's plain to see the value it's actually generating.

>>The premise of this post seems to be that material errors in Deep Research results negate the value of the product

No it’s not. It’s that it’s oversold from a marketing perspective and comes with some big caveats.

But it does talk about big time savings for the right contexts.

Emphasis from the article:

“these things are useful”

I'm just realizing this might finally be something that helps me get past analysis paralysis I have before committing to so many decisions online. I always feel like without doing my research, I'll get scammed. Maybe this will help give me a bit more confidence

On the flipside, you might end up getting scammed even worse because of incorrect analysis. For example if ChatGPT hallucinates some data/features through faulty research then you might be surprised when you actually make the decision.

While this will undoubtedly happen, I don't understand why this is a new phenomenon, the internet is filled with data with questionable accuracy. One should always be validating/verifying information even if Deep Research put it together.

I think the difference with Deep Research – and other hallucination and extrapolation-prone research agents – is that without assistance, that verifying synthesized information is much more of a slog than, say, doing your own research and judging the quality of sources as you go, which "deduplicates" querying and verifying.

Of course there are straightforward ways, in terms of UX, to make verification orders of magnitude easier – i.e. inline citations – but TFA argues that OpenAI isn't quite there yet.

Ultimately, if a research agent requires us to verify significant AI synthesized-conclusions as TFA argues, I'd argue research agents actually haven't automated tricky and routine work that keep us thinking about our research at lower level than we would like.

From my experience (having hit the Deep Research quota), I wouldn’t use it to build data tables like the article did, but for qualitative or text-based research, it’s incredibly useful. Useful enough that I’d justify multiple accounts just to increase quota. People keep saying hallucinations but in the reports I have built I have not noticed it to be a problem, again I am not doing quantitative analysis with it.

Yeah, probably true. But if it includes links and sources, at the very least it'll save me some time. I can cross-check faster than I can start the research

I have found it to be exactly this in a lot of cases. It helps answer or synthesize the data that answers questions I had that are good to know but not critical for me to understand.

It's what one imagines the first cars were like - if you were mechanically inclined, awesome. If not, screwed. If you know LLMs and how a basic RAG pipeline works, deep research is wonderful. If not, screwed.

I can't help but feel that it's different if a car runs 90% of the time but breaks down 10% of the time, and if it turns the direction you tell it 90% of the time, but the opposite direction 10% of the time.

Also that you won't necessarily know when it makes that wrong turn until it's too late (you're in the river now).

If you can't critically read the output of an LLM, you shouldn't be using it for the given task. Many people have made the (good) analogy to an intern.

Reply on news.ycombinator.com