Elicit

Two days ago, I helped my sister write her research proposal 📄. To be honest, it has been years since the last time I worked on academic research, and I can only remember how painful it was 😅. BUT …. GenAI products are game-changers. we didn't spend hours narrowing down a million articles like before, we used this product -> "Elicit" Which helped us find and analyze academic papers "efficiently"... it has screening tools and data extraction that are seriously powerful. Plus, you can organize everything with notebooks and libraries 📚. You can literally chat with the paper and pull out the key points in seconds, a very seamless user experience 💕. I also noticed that the output was relatively accurate compared to the current LLMs output, they attribute that to -> "factored verification" technique which detects and reduces hallucinations in AI-generated summaries of academic papers, this method involves breaking down summaries into smaller parts and verifying each part individually, making it easier for both models and humans to check for accuracy, I added an article about “factored verification” in the first comment if you wish to learn more about it 📑. Elicit ticked all the boxes for an AI product that gains user trust, Kudos! 🎉 #AI_product #Research #Product_review

🔍 🤵 I'm hiring a technical recruiter to help us find talent and make Elicit the best tool for research & reasoning. What you get: - Do something that matters. AI is probably the single biggest development in our lifetime. You can help apply it to society's most important decisions and questions by bringing the best minds to our team. - Your role scales with the company. We've grown from $0 to $1MM+ revenue within 4 months of raising our $9 million seed round, and are expecting to grow a lot more. - Competitive salary, initially as a 6-month contract. If it’s a good fit option to join full-time at the end. As a technical recruiter, you will work closely with me, James Brady, and Sarah Park to identify, attract, and hire top-tier engineering talent. You'll be a core part of building a product that 200k+ researchers use every month, shipping new features every week. As we work together, I'll introduce you to other people and organizations in the AI space so you can build your network. What we're looking for: - Technical - you know how to tell apart engineering competence at the high end - Scrappy - you want to join us in the whole cycle from sourcing to closing - Curious - you’re deeply interested in AI and our vision, and excited to talk to potential hires about it - Local, or willing to move - you're excited to hang out with us in our office in Oakland, and organize recruiting events in the area - Experienced - you have prior experience recruiting in-house (not agency-only or first-time) The best person has a strong network in tech, can communicate on different levels of complexity, and has a deep understanding of the (rapidly changing) technical landscape. Experience with recruiting for roles across the entire stack is a plus, including Kubernetes, Node, Python, NextJS, Typescript, and React. Want to work together? Send me your resume and a little about what excites you about Elicit and this role ([email protected]).

Did you know there are 14 advanced filtering options you can access through Elicit's search? This means you can specifically search for papers:  1. Published by a specific author 2. Published within a certain journal 3. Published in a certain year 4. That are open access  5. With at least a specific number of citations or references and many more! We created these filter commands to help make our semantic search more controllable, so you are always getting relevant, useful results. Bookmark this page that lists all available filter commands, and let us know if you give it a try! https://lnkd.in/g45CGJ-n 

We’ve completely overhauled Elicit’s semantic search! Quality has almost doubled in some cases - Elicit search results will now contain 78% more relevant papers, if you are searching for papers for a systematic review. Across all Elicit searches, you will find 40% more relevant papers. These results are based on systematic evaluation over 5 million papers. We're also launching advanced search using 14 filters. You can search based on title, journal, author, keywords, PubMed ID, DOI, publication year, open access status, citation count, and more. You can combine criteria to run AND / OR searches. For the full list of filters and how you can use them to control your search, check out this post: https://lnkd.in/g45CGJ-n And kudos to Justin Reppert, Ben Rachbach, and Adrian Smith for leading this massive project!

Our ML Reading List on Github contains 140+ papers compiled and organized by the Elicit team. It's become an invaluable source for new (and potential new) team members gaining a background in machine learning. Keep up with the reading list here: https://lnkd.in/gCnbPf3h

We added a filter for journal quality to help you search only top-ranked journals any time you want. You can simply move this new quality slider over to gradually narrow in on higher ranked journals, quartile by quartile. This was one of our top requested features, so we're especially excited to get this in front of users this week! ☝️ ICYMI, we launch a new user-facing feature every single week to keep us moving quickly in the right direction. You can read more about this practice and see what it'd be like to be part of our team here: https://lnkd.in/gmQC4whh

I've been confused by level frameworks - they often read to me as: "incremental steps towards becoming a manager" "just do bigger projects over time" "your level is years of experience" - not tied to fundamental value. Things got clearer once I started thinking about how levels add value. The simple polynomial model of growth: - Junior work adds constant value per contribution. - Mid-level work keeps on giving after it’s done - Senior work multiplies the value produced by your team - Principal work sets up feedback loops that lead to compounding growth This helped me understand why senior people are sometimes less effective at startups - if your leveraged contribution multiplies 0 value or 0 growth, you’re still at 0. I collected tons of examples of what our successes and failures Elicit look like for different levels of work here: https://lnkd.in/gWmmfzK2

One of our primary goals is to make quality research faster while exceeding human levels of accuracy, so we're especially excited about some recent quality improvements to Elicit: - Data extraction in columns without high-accuracy enabled is now 15% more accurate - Using high-accuracy mode to extract data from tables within papers is now 21% more accurate! You can stay up to date about these and other improvements to Elicit at our changelog: https://lnkd.in/g77TxaAr

Our Head of Engineering James Brady & internal journalist Adam Wiggins went on the Latent Space Podcast to talk about how we hire for AI engineers. He discusses the surprising chaotic qualities of language models and the role how pre-AI chaos engineering skills play in your success. If you're looking for or hiring for a similar role, watch the full episode and check out the linked template resources. They dive into: - Defining the hiring process - Defensive AI engineering as a chaotic medium - Tech choices for defensive AI engineering - How to interview for defensive AI engineering - Does model shadowing work? - Is it too early to standardize tech stacks? - Required knowledge for AI engineers - ML-first mindset - AI engineers and creativity - Sourcing AI engineers

You might spot a useful update we recently made to Elicit 👀 When reviewing supporting quotes that Elicit highlights for you during data extractions, quotes will now be ordered according to how relevant they are to your column instructions, with the first quote being the most relevant. We hope this helps speed up data verification and review for you!