YouTube Video IBsBixGTh-I  ## Metadata - Author: [[YouTube]] - Full Title: YouTube Video IBsBixGTh-I - Category: #articles - Document Tags: [[ai]] [[h1]] [[research]] [[xref]] - Summary: YouTube Video IBsBixGTh-I - URL: https://www.youtube.com/watch?v=IBsBixGTh-I You've asked for it, so here it is. Hereare my top AI tools for academia andresearch. So here are all of thedifferent sections I think an AI toolcan help with in academic life. We'vegot search, we've got mapping, we've gotreading papers, and reading papers isreally about reading and summarizing papers sothat it's easy to digest. Then we've gotmulti-document chat, and that's reallycoming in 2024. Then we've gotwriting and editing, and also workingwith data. So here are the AI tools thatI think you should use if you justdon't want to think about searching andall the rest of the AI universe. Here they are. Okay, the firsttool that I think you should use forsearching is Elicit. Here it is; it'sreally simple. I find it really powerfulfor my research field, but I know somepeople have a little bit of an issuewhen it comes to what databases they areusing. But nonetheless, this is the one Iwould use if I really had to decide onone tool. Elicit is brilliant because youcan ask a question, and once you click on it,you get all of this stuff. You get asummary of the top papers; if you pay, youget more papers, and then you get this. Nice little kind of summary of all ofthe different papers that it found.That's really great! It also gives youthe ability to get a quicksnapshot of that world that you've justentered with your semantic searchquestion. Delve into the deep stuff withthis right away. I absolutely love it!Here we go, and then we can add a newstep and so on and so forth. So that ismy first one; I really like it. I alsohave a really soft spot, a little softspot, like a little spot on me, that'ssoft. Ooh, that's soft! Um, it's SizeSpace. Size Space does a load of things,but if you're using it for search, I think it also does a really great job. Itdoes so much more than search! We'll betalking about that in a minute as itfits into other categories. But here youcan see that, yeah, you just click on hereand it gives you all of the papersthat you need to know about. So onceagain, we've got an answer from the topfive papers, and we get this very similarlayout to elicit, where you get a table,and you can also go all the way downhere and add your own column, and thenit's filled out with AI based on thatpaper. Does it get any better than that? I really like Sipas, and I think Space willbecome a tool that nearly everyresearcher will use at some point intheir career because it's so great. Okay,and then the last sort of tool that Iuse if I just want a yes or no or Iwant to know what the consensus is. Oh, Igave it away a little bit; here it is:Consensus. Consensus is really greatbecause you just ask a simple question,and you get a yes or no answer. So, dobeards make a man moreattractive? Okay, and then it will go awayand it will have a look at the research,and it will give you a load of differentthings. It will give you a oh, yes, no,40%. Oh, what do you think? Are beardsmore attractive? I think they are! Right then. So here we go; you got theconsensus meter, um, and you've got all ofthe different kinds of papers down herethat, you know, it referenced. So thoseare where I would go if I wanted to aska question or a semantic search questionto the literature. What would you use? Thenext section I think is very important,is mapping. Now, mapping does sort of likeseem optional sometimes, but I thinkmapping the literature really gives youthe ability to see if you've missed any. I really like spots, and my onlyone that I really go to, oh I got tosneeze. Keep pushing there until I don'tsneeze. Um, is Lip Maps. I really like LipMaps. So if I log into Lip Maps, you cansee here that I've got the seed paper,and then we're going to end up with allof this. I think we're going to—let'shave a little exploration. Articles, how do Iactually create a seed? Let'sclick this. This is the litmap here. So here we are; this is what Ilike about Lip Maps, and you can get thisum, completely customized to what youwant. Whether you want site count,references, you've got the Y AIS as well. Refcount, so all of this means it'scompletely customizable. It's the mostcustomizable and the most powerful I'veseen from all of the mapping tools that I've used. But if you do want somethingfree and you want to kind of like reallyuh, delve into research in a really kindof um, haphazard way, I guess thestructure is lacking in this app. But Ido use it regularly if I just want toexplore ideas, and that is ResearchRabbit. I really like Research Rabbit,because you can import your Zoterocollection, and then you just sort ofLike vomit stuff out theside, and here we go. You can see thatokay, suggested authors, and then you getthis, and you say, "Okay, I want to knowabout this person." Oh, look at thatcollaborators. Oh, it just keeps going andgoing. You cannot stop it from vomitingout information, and I kind of like that. At first, I was like, "No, I don’t like thisat all," but now I’ve kind of warmed up tothe idea that you can kind of explore inthis way. In terms of mapping, it doesn’tdo sort of like too great; that’s where Ithink LP Maps will do a lot better. Butin terms of just discovery from a mapkind of like interface, yeah, why not? Sothose two are my favorites at the moment.Which ones do you like the best? Okay, thenext one we got to talk about is readingpapers. Reading papers is really, um, aboutsummarizing papers quickly so you canunderstand them. In the early stage ofthe AI world, it was really about uploading a PDF and then sort of likeasking questions, and I still do that. ButI use ChatGPT, so nothing veryinteresting there. But there is somethingon, uh, SI space that I think is reallycool. You know, we’ve got this chat downhere, chat with paper. It's completelyFree, um, and you, yeah, ask a question. Oh, Iwant to know a question about that. Um,give me asummary, and this is kind of really howit works. It's a really simpleinterface; you chat with the paper. Absolutely simple and easy, and the onething I like about this is, obviously, notonly does it give you the summary, butalso, if you want to do that to a load ofpapers, you can add your own column here.Um, overall, it's just such a great way tointeract with papers, get a really goodsort of broad snapshot and a quicksummary. Clearly, you need to read them ifyou're going to cite them in the future,but ultimately, this is what I use. Um, ifI just want to sort of get across aload of literature, and then I want todelve a little bit deeper. They've even put this in here: playpodcast; you click here, and it'll giveyou a little bit of a podcast about thatpaper of climate change on globalbiodiversity. Let's dive in. I'm eager tohear this. What are the main threatshighlighted in this? GL warming, primarilydriven by human activities like fossilfuel burning and deforestation. There weare, so it turns all these papers into aPodcast really, I like it. It's just likeNotebook LM. We'll talk about that in amoment. Which one do you like? Forwhich one was this? Reading and summarizing papers, Idefinitely remembered I didn't need tolook. By the way, if you want to know howto use all of these AI apps and more foracademic writing, go check out my course.It's currently being updated for 2025,and I think that will be super valuablefor you going into the new year. Go checkit out. Alright, then, multidocument chat. This is where the AI world has reallycome on in the last year. I think it wasreally, really hard for them about twoyears ago to take all this information,vectorize it, and then make it searchable with asemantic question. But now this is where I end upwith this Notebook LM. I absolutelylove Notebook LM. You can upload up to50 sources. I really like that, and youcan upload Google Drive links, and youknow, you do have to be a bit carefulbecause it's completely free, which meansthat you could be, or you are, the product in some way.So I wouldn't uploadanything sort of like super sensitiveto this network or this platform. One thing that is really powerful isthat it's completely free. You can upload up to 50sources and then you can ask questionsabout those sources. You can see downhere that I've got nine sources. You can starttyping and it will give you a briefoverview of those documents. I reallylike it. It's completely free, and it hasother powerful features like theability to add a podcast of all ofthis stuff. You’ve also gotinteractive mode, with notesthat you can create down here. You know, go check out my other videowhere I talk about Notebook LM. Butultimately, if you’re thinking abouttalking about multiple documents, this isthe easiest way at the moment, I think,for most research purposes. Fifty documentsis all you’re really going to need andit’s going to really help you summarizewhat’s in those documents in asimple and easily accessible way. Check itout. The next thing we’re going to lookat is writing and editing. Writing andediting was sort of my leastfavorite activity during all ofmy research, but with the new AI tools,it is going to be better than ever andeasier than ever to create those perfectPros, okay, here we are. So, I just want to behonest with you; ChatGPT, Canvas, Iabsolutely love what they've done with this upgrade. So, here you can see I'veput a literature review aboutbeards, and then it's given me thisoutline. The one thing I love aboutCanvas is that you can go in; you canactually sort of write with it, andthen you can highlight things. Youcan ask ChatGPT, and you can say, "Youknow, expand."It will go through, and it willexpand that section there. This is onething I really like about it; you can seeit only changes that little bit. I'mpointing to the screen, but that doesn'thelp you, does it? You can see it onlychanges that little bit, and you can useit to work on that, um, sort of like, uh,growth of the ideas in your mind intoactual paragraphs that someone will read. But I know that is not how everyonelikes to work. Some people actuallylike to use an actual AI writer foracademia and research, and that is Jenny.Jenny, I think, is my favorite one at themoment. There are other ones like Yomuand stuff, but Jenny has just been able todo this very well. Now, older academicsMany like myself will look at Jenny and go,"It's too confusing; I don't like it." But Ifeel like the next generation ofacademics will be writing in this way,and that's because older academics likemyself have a workflow; we've got ourfirst draft, and then we add citations. Then we go through and check it —blah, blah, blah. But in Jenny, you kind ofdo all of that, as you go, which isa little bit sort of counterintuitive for someone who'swritten papers like I used to write them.But I think newer generations ofacademics will absolutely love Jenny,because you can go in and obviously juststart typing and creating.Information will spill out what youneed. I think it is just a veryuseful tool that you can use,so we can accept that. And yeah, youknow, that clearly wasn't good; it was justsmashing the keyboard. But yeah, wecan go up here; we can choose all of thedocument settings. I absolutely love it.In Jenny, you can add citations now fromyour own library and also what it foundonline. So overall, I think Jenny is agreat AI writing tool if you are so inclined;but if not, I love ChatGPT. GPT Canvas feels like agenuine collaboration. All right, whichone's your favorite for writing andediting? Let me know. The last thingthat we've got to talk about is data andanalysis. So here we've got data. Um, yeah,that's why I just wrote data. I don'tknow why. Look, we've got data, but this iswhat I really like: Julius AI. JuliusAI has been a powerhouse. I really likeit. It's like having a little, um, dataanalysis person in your pocket that’slike, "Do this." "This is better for you." "Doit over here." All of that stuff, so Ireally love it. Um, statistics andanalysis was never something that Iwas super interested in during my PhD,but now I think you can upload data. Youcan explore that data with semanticsearch; you can get visualizations. Youcan get all of the graphs you need forthat first interrogation of any raw data. It works with a load of sort data. Itoutputs Python code so you can run it onwherever you want to run it. It isabsolutely brilliant. I really recommendyou check out Julius AI. But of course,some people are like, "Oh, but I’ve gotsensitive data; I don’t want to upload itto the internet." I completely get it. NoWorries in that case, I run a localversion and it's called DataLine. Ithink it's really great. It does take alittle bit of time to kind of get usedto how it works and get it working inyour local environment, butthis is what it looks like, and you cansee here I can add a COVID connection. Then this is some data that Ican talk to, and you know, I don'tknow, give me a summary of the data,and then it will generate results andgenerate graphs or generate sort of likea response to that data. I think it'sreally great if you want to keep all ofthat information on your desktop. It doesuse an API from OpenAI, so that'sthe ChatGPT makers. But ultimately, hereis just a really great way to keep allof that data on your computer, making sureit doesn't go to a nefarious third party,but it stays on your computer. I absolutely love it. In fact,I was given a talk at Harvard this last year,and they actually havetheir own sandbox. So if you do havea sandbox, you probably don't needsomething like this for your data, andthe sandbox will keep all of that datawithin your institution. Servers, so if you've got that as part ofyour university, great. If not, somethinglike this may be able to help you. Oh,I love that! All right, then let me know inthe comments what you think. If you wantto know more about AI tools, check outthis video from last year where I talk aboutall of the best free AI tools foracademia and research. I think there'llbe something in there that you'll love.Go check it out., YouTube Video IBsBixGTh-I, YouTube, 01kac33wxty0azvz62tdc2ef6p.