OpenAI Releases Prism, a Claude Code-Like App For Scientific Research (engadget.com) 15
OpenAI has launched Prism, a free scientific research app that aims to do for scientific writing what coding agents did for programming. Engadget reports: Prism builds on Crixet, a cloud-based LaTeX platform the company is announcing it acquired today. For the uninitiated, LaTeX is a typesetting system for formatting scientific documents and journals. Nearly the entire scientific community relies on LaTeX, but it can make some tasks, such as drawing diagrams through TikZ commands, time-consuming to do. Beyond that, LaTeX is just one of the software tools a scientist might turn to when preparing to publish their research.
That's where Prism comes into the picture. Like Crixet before it, the app offers robust LaTeX editing and a built-in AI assistant. Where previously it was Crixet's own Chirp agent, now it's GPT-5.2 Thinking. OpenAI's model can help with more than just formatting journals -- in a press demo, an OpenAI employee used it to find and incorporate scientific literature that was relevant to the paper they were working on, with GPT-5.2 automating the process of writing the bibliography. [...] Later in the same demo, the OpenAI employee used Prism to generate a lesson plan for a graduate course on general relativity, as well as a set of problems for students to solve. OpenAI envisions these features helping scientists and professors spend less time on the more tedious tasks in their professions.
That's where Prism comes into the picture. Like Crixet before it, the app offers robust LaTeX editing and a built-in AI assistant. Where previously it was Crixet's own Chirp agent, now it's GPT-5.2 Thinking. OpenAI's model can help with more than just formatting journals -- in a press demo, an OpenAI employee used it to find and incorporate scientific literature that was relevant to the paper they were working on, with GPT-5.2 automating the process of writing the bibliography. [...] Later in the same demo, the OpenAI employee used Prism to generate a lesson plan for a graduate course on general relativity, as well as a set of problems for students to solve. OpenAI envisions these features helping scientists and professors spend less time on the more tedious tasks in their professions.
Finally! (Score:2)
Where was this all my life? This is actually something AI can do extremely well!
Re: Finally! (Score:3)
Hello fellow low UID o.g.!
Re: (Score:2)
And first post, too. Here's to us keeping the dream alive :).
So it makes researchers 20% slower? (Score:2)
Because that is what it essentially did for any real programming besides simplistic boilerplate: https://mikelovesrobots.substa... [substack.com]
Re: (Score:3)
I don't consider using but I can see the use case. "Convert this image to drawing commands" could be useful to some. A drawing in a paper won't need code structure or maintenance, so code maintainability is no concern. Also editing tables in latex is tedious and tricky, and is a table at which AI should excel (generating templates, reformatting).
Re: (Score:2)
Also editing tables in latex is tedious and tricky, and is a table at which AI should excel (generating templates, reformatting).
Not that tricky, really. If this thing is essentially just a LaTeX editor addon, it will not do much.
Re: (Score:2)
There are also a lot of use-cases in restructuring documents. You decide to replace one LaTeX package by another? Now you need to change that command in subtle ways, but some replace-all will break havoc. If (and only if) a model understands enough LaTeX that would be a good task for a LLM. Another issue is, that LaTeX is token heavy. With 1 million context in GPT that is no large problem, but for smaller local models you're filling the context quite fast. Personally I was missing some good integration with
Re: (Score:2)
Because that is what it essentially did for any real programming besides simplistic boilerplate: https://mikelovesrobots.substa... [substack.com]
Actually, in research capacity, provided it backs up its "conclusions" with links to the original documents, it can help speed things up. it's about the only thing current generation AI is good at. Granted, some goober is going to suggest you can pose a research question to the AI and have a fully written and cited report without any domain-knowledge human involved, and that person is going to need to be slapped, hard, repeatedly, until they sit down and shut up. There are way too many people that think cur
Re: (Score:2)
Indeed. Well, "somewhat better search, but it hallucinates on occasion" is the one thing LLMs actually can do. You will still generate a low quality publication unless you invest a lot of time yourself to verify everything.
Admittedly, there are a lot of low-quality publications out there, even with supposedly "high quality" conferences and journals. Sturgeon's law is very true for scientific publications. When I did peer-reviews, I several times got editor feedback along the lines of "you were the only one
cloud based latex copy? (Score:2)
Who needs it, CERN already pays for my unterleaf.
Isn't Prism a government surveillance name? (Score:2)
Your data, exfiltrated to the AI companies. (Score:2)
that's Ok though, you can now work 50, er no, 10, er wait -20 percent faster now!
More junk papers (Score:2)
This will no doubt make it easy for fraudsters to compose junk science papers.
Just give its AI a prompt such as: "Write a science paper about how seed oils are bad for my health."
It will quickly produce a fake paper with references that will appear legit at some level of inspection.