tom_hartke

5 ms·

1.
by tom_hartke 2y ago
A decent fraction of research is open access these days (it's highly dependent on the field, though). That fraction is also growing. We're building systems to access full texts whenever possible.
2.
by tom_hartke 2y ago
That makes sense. You should be able to right click - save as PDF and it will preserve the links, etc.
3.
by tom_hartke 2y ago
We'd love to talk and see if we can provide this. Seems like it would be really helpful. Can you email us (support@undermind.ai)?
4.
by tom_hartke 2y ago
Hadn't seen that. Very complementary. We don't address staying up to date or pull in community value metrics (other than total citations). It's a somewhat different goal to broadly gather emerging ideas and stay informed.
5.
by tom_hartke 2y ago
Can you be more specific? Like break it out into AND and OR statements? Or just more iteration back and forth? We find people more familiar with the system learn better strategies than the LLM can suggest.
6.
by tom_hartke 2y ago
If it's easy/fast to check the literature, hopefully people's instinct changes. If we had time, we should be doing that for any small idea.
7.
by tom_hartke 2y ago
Just save the report url! It will persist. Also, on the usual system if you're logged in (instead of the HN free link), all your searches automatically get saved on a "history" page.
8.
by tom_hartke 2y ago
I think we associate learning/discovery with those moments because they happen together, not because they're causally related. I think this is somewhat equivalent to how we used to have to learn 100 different integrals and derivat
9.
by tom_hartke 2y ago
Thanks! That's helpful to hear. Honestly just did numbers because the LLM has no trouble remembering which is which, and it's easier to programmatically parse out the citations to build hyperlinks (compared to names/years, wh
10.
by tom_hartke 2y ago
Emphasis is meant to be on "mind", like supporting better reasoning, but point taken.
11.
by tom_hartke 2y ago
We removed the signup requirement for HN today - try this link: https://www.undermind.ai/query_app/promotion/
12.
by tom_hartke 2y ago
Not at the moment -- we're currently searching the abstracts of most major journals (which are public even for paywalled papers) which have been compiled in the Semantic Scholar database ( https://www.semanticscholar.org&#x2F
13.
by tom_hartke 2y ago
Not a bad idea. Want to avoid too much complexity in pricing, though. Decision fatigue.
14.
by tom_hartke 2y ago
Latency, compute required, and lack of full texts (paywalled publisher content).
15.
by tom_hartke 2y ago
In short, yes, though it's geared toward topic search. From a strategy perspective, we designed it for topic search because it makes more sense to find everything on a topic first, then filter for the most recent, if recent is what y
16.
by tom_hartke 2y ago
The few minute time delay is primarily because of the sequential LLM processing steps by high quality LLMs, not database access times. The system reads and generates paragraphs about papers, then compares them, and we have to use the highes
17.
by tom_hartke 2y ago
Ours is slow, but accurate, even for complex topics. The rest are fast, but generally can't handle complex topics. (There's more nuanced explanations in other comments)
18.
by tom_hartke 2y ago
It's worth highlighting that first result is exactly what you asked for, given all 4 of your criteria: 1. It's on adults. 2. It's longitudinal over multiple years. 3. It studies variations in brain volume. 4. It focuses on
19.
by tom_hartke 2y ago
Semantic Scholar seems more focused on 1. being the data provider/aggregator for the research community, and 2. long term, I think they plan to develop software at the reading interface that learns as a researcher uses it to browse p
20.
by tom_hartke 2y ago
I think the biggest difference is our focus on search quality, and being willing to spend a lot on compute to do it, while they focus on systematic extraction of data from existing sources and on being fast. It's a bit of an oversimpli
21.
by tom_hartke 2y ago
Potentially. Given the latency and the cost/compute we put into each result, it doesn't fit the usual API mechanics. What use case are you thinking of?
22.
by tom_hartke 2y ago
Agreed, exa is great - particularly, it's the best thing I've found for fast web retrieval of slightly more complex topics than Perplexity, Google, etc can handle.
23.
by tom_hartke 2y ago
I ran these two example searches we have on our homepage on AnswerThis: (3D ion shuttling) https://undermind.ai/query_app/display_one_search/b3767fb7b6... (laser cooling to BEC) https://undermind.ai&#x2
24.
by tom_hartke 2y ago
For a meta-analysis, you might want to try the "extend" feature. It sends the agent to gather more papers (we only analyze 100 carefully initially), so if your report might say "only 55% discovered", could be useful. (Al
25.
by tom_hartke 2y ago
You should be able to try it here without loggin in: https://www.undermind.ai/query_app/promotion/ (set up for HN today). If not message support@undermind.ai and I'll set you up.
26.
by tom_hartke 2y ago
A few possibilities: - We only use abstracts for now. Have to make sure you ask for something present there. - Did you ask for a scientific topic? (Sometimes people ask for papers by a specific author, journal, etc. The system isn't en
27.
by tom_hartke 2y ago
We're trying to bias the system toward more autonomous execution, rather than a "copilot"-like experience where you iterate back and forth with the system. That lets us run more useful subroutines in parallel in the backend,