Professor of Strategic Management. Bridging AI and Academia. AI news, tools, prompts, and workflows for researchers and educators
Europe
Joined February 2023
- Tweets11K
- Following1.4K
- Followers15.3K
- Likes4.7K
Pinned Tweet
AI tools to supercharge your research and teaching!
Well, it's about time I updated the list a bit. Not all of them are here. I will likely add some more, but I didn't want to delay it any longer.
The order in which they are written doesn't mean anything.
✨ SciSpace (@scispace ): Chat with papers, understand math, paraphrase text. 🔗 scispace.com
🐇 Research Rabbit (@RsrchRabbit ): Find & organize academic papers with AI-powered connections. 🔗 researchrabbit.ai
📝 Glasp (@_Glasp ): Highlight, take notes, and get AI summaries from web & PDFs. 🔗 glasp.co
🤖 Microsoft Copilot (@Copilot ): Your AI assistant for content creation, data analysis, and more. 🔗 copilot.microsoft.com
🧠 Gemini (@GeminiApp ): Search, generate text, brainstorm ideas, integrated with Google Workspace. Deep Research🔗 gemini.google.com
🔍 Semantic Scholar (@SemanticScholar ): AI-powered academic search engine for precise results. 🔗 semanticscholar.org
✍️ ChatGPT (@OpenAI ): Versatile language model for summarizing, rewriting, and generating text, editing code and much more. Deep Research 🔗 chat.com
💡 NotebookLM (@NotebookLM): AI research assistant grounded in your uploaded documents (also, creates podcasts from papers and mindmaps). 🔗 notebooklm.google/
🚀 Grok3 (@xAI): Advanced AI model with strong reasoning and real-time web analysis (access via X). Deep Research🔗 grok.com/
🦾 Manus AI (@manusai ): Autonomous AI agent for complex task execution like research and content creation. 🔗 manus.im/app
☁️ Claude (@AnthropicAI ): AI assistant for summarizing papers, coding, and creative tasks. 🔗 claude.ai
🤝 Consensus (@ConsensusAI): AI search engine that synthesizes findings from academic research. 🔗 consensus.app
🔎 Elicit (@elicit_ai): AI research assistant for finding and summarizing relevant papers. 🔗 elicit.com
🔬 Scite (@scite_): Checks citation context and analyzes the credibility of sources. 🔗 scite.ai
💡 Perplexity AI (@perplexity_ai): AI-powered search engine with citations from credible sources. 🔗 perplexity.ai
🌐 DeepL (@DeepLcom ): AI-powered translation (31 languages) and rewriting tool. 🔗 deepl.com
🎙️ Gladia (@gladia_io ): Accurate and multilingual audio transcription API. 🔗 gladia.io
🖋️ Paperpal (@teampaperpal ): AI grammar checker and academic writing tool. 🔗 paperpal.com
📜 Scholarcy (@scholarcy ): Online article summarizer tool using AI. 🔗 scholarcy.com
🔗 Connected Papers (@ConnectedPapers ): Visual tool for discovering and exploring academic papers. 🔗 connectedpapers.com
🖊️editGPT (@editgpt ): AI-Powered Editing, Grammar and spell check, style suggestions, tone adjustments, custom prompting, ChatGPT integration. 🔗 editgpt.app/
🎶Macwhisper (@macwhisperapp ) : AI-powered dictation, audio transcription, on-device processing, drag-and-drop, automatic meeting recording, multiple language support, various export formats. 🔗 goodsnooze.gumroad.com/l/mac…
🎶Superwhisper (@superwhisper ): AI-Powered Dictation, offline functionality. 🔗 superwhisper.com/
🖊️GrammarlyAI (@Grammarly ): Writing Assistant, Grammar and spell check, style suggestions, tone adjustments. 🔗 grammarly.com/
🖼️Canva (@canva ): Graphic Design, AI-powered design tools (Magic Studio), collaboration features, print services, Canva for Education. 🔗 canva.com/
I got annoyed with Claude because it’d written just a bad draft message for my students.
The second version was much better.
I still prefer my own messages.
No doubt.
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
I’m liking Hark so far. I haven’t used it much yet, but my first impressions are good. Have any of you tried it?
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
I can imagine authors rubbing their hands together at all the possibilities AI offers for research. And at the same time, I can imagine reviewers doing exactly the same, given how much easier it is for them to criticize papers now.
Competitive parity?
I can imagine authors rubbing their hands together at all the possibilities AI offers for research. And at the same time, I can imagine reviewers doing exactly the same, given how much easier it is for them to criticize papers now.
Competitive parity?
I’m liking Hark so far. I haven’t used it much yet, but my first impressions are good. Have any of you tried it?
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
Are you finding it harder to publish now than before generative AI came along?
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
Ups. Human reviews worse than AI reviews? 😬
Massive. New Harvard study finds that AI-written paper reviews now beat the best human peer reviews on every measure the authors tested.
Peer review is slow and often misses errors. Earlier studies concluded AI should only assist human reviewers, not replace them.
They used 134 reviews from rapidPeer, a Harvard service that pays expert reviewers well. Claude Opus 4.8 reviewed the same 1124 papers, and GPT-5.5 plus 3 former journal editors judged the pairs blind.
GPT-5.5 picked the AI review in all 132 decided pairs, and the editors picked it 87% of the time. Only 5.5% of the AI's claims about the paper were wrong, versus 15.9% for the human reviews.
Teams that depend on expert critique can now let AI do the first full review pass, and keep a human for the final decision.
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
Last Wednesday I gave a talk at USC's Econ department on "Skills in Economics Research".
I turned my talk into a YouTube video teaching the use of skills, and the 7 mistakes I see economists consistently making with skills.
Here are the 7 mistakes:
1. Economists often don't make their own skills. Instead they look for libraries of skills made by other people. The skills they find only partially fit their use case. I show in the video that skills are typically very easy to make.
2. They make too few skills - often countable just on one hand. I have hundreds of skills across my projects. I suggest economists adopt a "rule of 3" with respect to skills: if you notice yourself doing the same thing three times, make a skill.
3. They make unnecessary skills. Often economists without significant pre-AI coding experience will make skills in situations a deterministic script is more appropriate. I explain through a case study of economics research on the "cereal industry" how one can understand whether their current problem is "skill-shaped".
4. They don't use their previous agent conversations to improve their skills. Both Codex and Claude Code store session transcripts on your computer. Your agent can use these to improve existing skills and identify useful new ones—I show how.
5. They don't clean up their skills. As projects evolve or new models get released, skills often "go stale", and start to work against you - costing you tokens and time rather than saving them. I explain a low effort mechanism to do a "spring cleaning" of your skills.
6. They don't use skills for project management. My two most useful skills are the skill to handle all my meeting transcripts and the one to handle my email. Especially for senior faculty who are not writing code, but are managing many projects - these project management skills are invaluable (see @cblatts' Claude Blattman). I show how to set up and use these skills.
7. They don't compose their skills. Skills can call other skills. For the ambiguous knowledge work which characterizes much of economics research, it's important to think carefully about whether a repeatable workflow makes most sense as one skill or several skills working together. I provide guidance on how to make that decision, with examples.
If you've been following me for a while, this is without a doubt the best video I've made, and will hopefully be a lower bound for what is to come!
Full video link:
youtu.be/BQahybA5lMg
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
You can now run pplx-decider-flash-v1.2 locally on your Mac, without an API key.
Decider for Mac is an app that runs the 4B decision model on device. Ask a question about text or an image and get a probability back, all on your machine.
pplx.ai/decider-for-mac
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
Ups. Human reviews worse than AI reviews? 😬
Massive. New Harvard study finds that AI-written paper reviews now beat the best human peer reviews on every measure the authors tested.
Peer review is slow and often misses errors. Earlier studies concluded AI should only assist human reviewers, not replace them.
They used 134 reviews from rapidPeer, a Harvard service that pays expert reviewers well. Claude Opus 4.8 reviewed the same 1124 papers, and GPT-5.5 plus 3 former journal editors judged the pairs blind.
GPT-5.5 picked the AI review in all 132 decided pairs, and the editors picked it 87% of the time. Only 5.5% of the AI's claims about the paper were wrong, versus 15.9% for the human reviews.
Teams that depend on expert critique can now let AI do the first full review pass, and keep a human for the final decision.
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
I welcome this update to the Guidelines on the responsible use of Generative AI in research, but we still need other type of initiatives from the EU
Link: research-and-innovation.ec.e…
AI Research Tools 👨🎓 🧬🧪 🐬 🔬 retweeted
Gen AI creates a kind of cognitive disconnection from the things we write if we use it extensively.
And I do not mind disconnecting from certain emails or certain administrative tasks, but I love my job and the pleasure of writing a paper.
That feeling of owning what you write ... I miss it if I don't control the process and I do not want to lose it.
Gen AI creates a kind of cognitive disconnection from the things we write if we use it extensively.
And I do not mind disconnecting from certain emails or certain administrative tasks, but I love my job and the pleasure of writing a paper.
That feeling of owning what you write ... I miss it if I don't control the process and I do not want to lose it.
I welcome this update to the Guidelines on the responsible use of Generative AI in research, but we still need other type of initiatives from the EU
Link: research-and-innovation.ec.e…