The real game changer is the new "GPTs" feature in GPT4, where you can now configure your own GPT "personnality" with precise instructions of how it must interacts with you. With that, you get perfect consistancy with its answers. You can make a GPT for co-writing, one for data analysis based on your topics or specific needs, and one as a reading assistant (such as summarizing and etc...). Limitation made from the lack of "rules" and "chaotic behaviour" is now gone. And the options to create your own AI assistant are limitless. What a year. Thanks for your videos, always nice to get news from you about the world of academia!

@waleedmandourlecturer

6 ай бұрын

I second your thoughts. It would be great if there is a video for an AI Assistant for Academia on GPTs.

@kathrin9674

6 ай бұрын

how do you do it? I have GPT 4.0 and trained it really according to tutorials, but I starts hallzunating from time to time again:(

@MatthieuPrugne

6 ай бұрын

@@kathrin9674 Pretty much just crafting and crafting... trial and error. For example as a reading assistant, i gave him precise instructions on what to look for in the articles i provide him and how to summarize it in a constant format so i can just upload a pdf and it automatically takes out what i want. The "normal" GPT4 was doing the exact same thing but I always had to delimits my needs at every single chat i was doing. Now those "preface" rules are pre-established before asking what I need. There is gonna error for sure. You can't take what it gives you as it is. But there's some crazy time saving trick.

@pointingtojesus247

6 ай бұрын

Prompting is the secret. Here are some of my prompts that I shared with fellow PhDs. Checking for Coherence "Evaluate the coherence and clarity of my argument by identifying the thesis statement, main claims, and evidence in my paper draft. Are my points logically structured and substantiated with robust evidence?" Writing Assessment "Critically assess the quality of my writing by examining my use of language, tone, and academic conventions. How can I enhance the precision, conciseness, and persuasiveness of my text and argument?" Literature Review "Summarize the key themes and debates found in the literature on [specific subject], identifying major authors and their contributions to the field." Research Question Formulation "Based on the current gaps in the literature concerning [topic], what would be three compelling research questions for a thesis?" Methodology Development "Outline a comprehensive methodology for a study investigating [research question], including data collection and analysis techniques suited to this research design." Data Analysis "Explain how one would apply thematic analysis to qualitative data related to [topic]. Provide an example of coding and theme development." Critical Analysis "Critically analyze the findings of [specific study or article], discussing the strengths, limitations, and implications for the field of [discipline]." Theory Application "How would [specific theory] be applied to interpret the phenomenon of [specific case or situation] within the context of [discipline]?" Hypothesis Testing "What statistical methods would be appropriate for testing the hypothesis that [specific hypothesis], and how should the data be prepared?" Interdisciplinary Analysis "Propose an interdisciplinary framework that could integrate insights from [Discipline A] and [Discipline B] to investigate [common problem]." Ethical Consideration "Identify potential ethical dilemmas in researching [topic] and suggest strategies for addressing them in accordance with [specific ethical guidelines or standards]." Theory Critique "Provide a critique of [specific theory] in the context of [recent developments or findings] in the field of [discipline]." Publication Strategy "Develop a publication strategy for research findings on [topic], considering target journals, audience, and the potential impact of the research." ©Frank Onuh LinkedIn: Frank Onuh Twitter: thefrankonuh

@gloombla

6 ай бұрын

He has a video about it and it seems it’s still not that good

To be honest, I'd love to use these tools, however, I would want to run it locally. I am worried about data breaches.

@user-zm1gu6bj4k

6 ай бұрын

Try "quivr"

@AlgorithmicEchoes

6 ай бұрын

If you can do a bit of python, you can try dify. Completely locally deployable. You can when use your own language models.

Hi, Andy! I've got a couple of important questions for you. 1. Can you make such reviews of apps with the participation of the creator? Meaning can you get in touch with the guys who created Petal and make a review together - you doing your exploration of it and them explaining things and showing stuff. I think this way it would be very informative and more interesting. We will get to know more about the app and the creators will be given the chance to show it in its best light. 2. Can you make a video about how to use Chat gpt or other app locally for academic purposes? Can we use Petal or other app of the sort locally on our machine and feed it with all the data we want? I think that would make us more willing to really safely use it.

Great video! also loved the glitch at the end for making me laugh :D

Hey Andy, Thank you for this. I think this will be very helpful for my PhD journey., I have been relying on Scholar Assist for this kind of thing.

Great video as always, Andy! As a doctoral student, I find your channel extremely helpful. :) can you make a video about modified GPTs (new ChatGPT features) for research?

Hi Andy, love your videos, I listen to them every morning before I set out to the lab. I have a question for you. Do you think depending on AI for research assistance will end up getting biased? In the sense that in search engines, SOP modified entries would always get a preference. (Also I don't know computer science, I'm a biology student, so I'm really just curious about if this could be a problem)

@DrAndyStapleton

6 ай бұрын

A tool like this still relies on you going and finding articles from multiple sources. I think it will still rely on the Researcher using AI as only one way of finding information. AI helps streamline workflows - after using them for a while you realise that you still need to put effort into guiding them and getting the best from them.

@nalonda3319

6 ай бұрын

@@DrAndyStapleton I see what you mean, however given it has learning capabilities, if it goes as far as to give suggestions on future steps one might take, or suggest other papers, for newbies (and even easy going researchers) there is a very real chance to have a biased literature review. So with AI tools in research, how far would you feel is too far?

Sorry if I missed it but Is the gpt plus version still needed to use this plugin.

Hey, do you know what language model the AI is using? Because I have the feeling that GPT4 will still be superior

Dr S, I actually like writing...but my professors think I suck. Can you talk about the best tool for using AI to work as my editor? I write and it goes through and replaces my professor for edits. I can go through the edits and put stuff in my own words and then resubmit...

Petal or docanalyzer? Which one do you prefer?

I get following error while uploading my pdfs. "AI performance may be degraded for documents with more than 1,000,000 characters. Please consider splitting them into smaller sections if possible."

Would love to see if you would expand AI to museum research and the use and creation of datasets for AI use. TIA.

Mate, I'm so glad you got caught and ended up running this amazing channel. It helps to see another fellow scientist that's strives to do something impactful in spite of the current state of 🦇 🧹 Academia 🦇 👻

@kingsleymoore5427

6 ай бұрын

"I'm so glad you got caught" Yo bruh, hahaha 😂, like seriously

Cada hora é um programa novo...

similar to quivr and/or powerdrill ?

I really can't thank you enough! Alone kahubi you mentioned in a video before saved me TONS of time!

Until people improve the metadata and paradata, GPT-AI will be limited. Metadata is improved by AI companies paying for human reinforced learning, but paradata is harder, less understood and less focused. But AI runs on labels and tags and is only as good as those labels and tags how those labels and tags are created, and why.

Dr. Stapleton, I hate to say it, but I don't think the computer is lacking in its ability to answer questions, I think you just don't ask structurally good questions. What does "best" mean in the context of your research: most efficient, most conclusive data, best support to the hypothesis, perhaps it is the one with the most fluent language or grammar... you asked it a question that was too vague for it to answer logically without making some assumptions about your meaning.

Wait... did I just watch an 11 minute ad?

This one is the first AI I have seen that acts like a human. Most people, when you ask them hard questions, say “I don’t know”. This AI does that, too! Breakthrough. 😂

Hmmm looks like an RAG? Think researchrabbit has something similar too.

I find it incredibly ironic that many of the ex academics of KZread who are calling out their industry for a complete lack of ethics, turn around and make money off of undisclosed partnerships with sponsors, partner with sponsors like betterhelp, masterworks, and established titles (all scams, some actively dangerous) as if it's not unethical if the people you're harming don't have PhDs.

Hello dear scientist colleague. I am first!

An interesting tool... but are you using an affiliate link, because if you are must disclose it or KZread may slap you on the wrist, and if you are profiting from it you should declare this to your viewers. I like your videos, an unbiased view of academic tools, but if you're being financially incentivised I would like to know.

Hi Andy, you've probably noticed that ChatGPT has made an update which allows creating your own custom GPTs for a specific purpose, like assistants. It would be really interesting to see how you create a custom GPT for avengers academia and PhD research purposes. What do you think about that?

Humata is way more advanced.

Dude you're getting paid to shill this AI based on your affiliate link. Throw in some honesty here. You're an academic aren't ya?

@alextrebek5237

6 ай бұрын

Unlikely because he also covered claude and anthropic. I tested these with petal now. So far it seems like Claude explains much better, but claude has a MUCH lower context window and couldnt use large pdfs.

@shubhammishra7044

5 ай бұрын

Which one is best?

@shubhammishra7044

5 ай бұрын

What do you guys think about paperpal? I'm really confused, a lot of tools are available!

Sensation! AI can write your PhD! Oh, wait. Nevermind.

Hmmm.... Honesty, i am not happy with this video. Most of the times the service didnt do what you want, and at other times it executed instructions but incorrectly. Furthermore, any AI service can spit out responses, but the question is.... are the information correct? The few demonstrations of the service in this video have already shown me that it is useless. I need to be able to trust a service. Im using some services from your other videos, but I wont try out this one, sorry pal.

Skynet is laughing

I didn't like your video this time. It seems you are exploring the app in front of the camera for the 1st time. I would prefer you had prepared your speech more carefully before.

Anybody with knowledge in local LLMs could make this app. Don't fall for the trendy scams. Make it yourself.

•> I believe we are meant to be like Jesus in our hearts and not in our flesh. But be careful of AI, for it is just our flesh and that is it. It knows only things of the flesh (our fleshly desires) and cannot comprehend things of the spirit such as peace of heart (which comes from obeying God's Word). Whereas we are a spirit and we have a soul but live in the body (in the flesh). When you go to bed it is your flesh that sleeps but your spirit never sleeps (otherwise you have died physically) that is why you have dreams. More so, true love that endures and last is a thing of the heart (when I say 'heart', I mean 'spirit'). But fake love, pretentious love, love with expectations, love for classic reasons, love for material reasons and love for selfish reasons that is a thing of our flesh. In the beginning God said let us make man in our own image, according to our likeness. Take note, God is Spirit and God is Love. As Love He is the source of it. We also know that God is Omnipotent, for He creates out of nothing and He has no beginning and has no end. That means, our love is but a shadow of God's Love. True love looks around to see who is in need of your help, your smile, your possessions, your money, your strength, your quality time. Love forgives and forgets. Love wants for others what it wants for itself. Take note, love works in conjunction with other spiritual forces such as faith and patience. We should let the Word of God be the standard of our lives not AI. If not, God will let us face AI on our own and it will cast the truth down to the ground, it will be the cause of so much destruction like never seen before, it will deceive many and take many captive in order to enslave them into worshipping it and abiding in lawlessness. We can only destroy ourselves but with God all things are possible. God knows us better because He is our Creater and He knows our beginning and our end. Our prove text is taken from the book of John 5:31-44, 2 Thessalonians 2:1-12, Daniel 7-9, Revelation 13-15, Matthew 24-25 and Luke 21. Let us watch and pray... God bless you as you share this message to others.