Project Read

It was a joy to join my friends @ValContessePhD and @HollyLanePhD in Gainesville this week. We shared how are @ProjectReadAI and @UFLiteracy are narrowly tailoring AI to specific use cases that truly support teachers with some of the thorniest challenges in phonics instruction: 1) spelling analysis, 2) small group planning and 3) increasing/individualizing student practice

NEW FREEBIE 🚨 We teamed up with @projectreadai to bring you the Gentle Cindy Notebook — a 50-page printable full of engaging activities to help students practice the Gentle Cindy rule! What’s inside: ✨ 3 decodable passages (short, standard, and long) ⏱ A 1-minute ORF fluency passage + teacher evaluation sheet 🧠 Comprehension questions 🎲 Roll & Read game 🎯 Gentle Cindy Bingo 🖼 Plus an IMSE Gentle Cindy poster Perfect for small groups, centers, or extra fluency practice. Grab the freebie and start building confidence with Gentle Cindy! 📖 imse.com/digital-resour… #IMSE #StructuredLiteracy #ScienceOfReading #PhonicsFun #TeacherFreebie

Thank you @UFLiteracy, @HollyLanePhD and @ValContessePhD for collaborating on a webinar this Wed at 7:30 pm ET. Supporting Readers at Home: Tools and Strategies for Remote Learning us06web.zoom.us/webinar/regist…

A message on where we stand, and how we are offering support to families and educators with students learning from home. tinyurl.com/projectreadsch… Please share with any families or educators you know whose children could benefit.

@HollyLanePhD This sounds familiar!!

This piece describes exactly what we recommend for UFLI instruction: 30 minutes of whole-class + 15 minutes of targeted small group (7-8 minutes for each of the previous week's concepts) based on weekly PM data. It works beautifully.

Last year I received an email from a teacher that I won't ever forget. I'm excited to bring you that story directly from the teacher who shared it, over a live conversation. Register here: lnkd.in/ei-YEkhb 2nd grade teacher Cristie Worthington shares the story of her classroom, and how Project Read's AI Tutor drove transformational literacy growth and self-confidence of her students.

@HollyLanePhD @BeCre8ive50 @vivramak @C_Hendrick Thank you @HollyLanePhD our team loves working with you all. @UFLiteracy is changing the game for teachers and students across the world.

@BeCre8ive50 @vivramak @C_Hendrick I've interacted with many developers, but Viv is the only one I've met who truly understands the things that are important for phonics instruction (content and methods). The lack of knowledge among the others has always become obvious with the design choices they have made.

There's a critical corollary here wrt AI x edtech, and it's not what most people are talking about. But it has major implications for designing with fidelity to what @C_Hendrick calls instructional invariants. TLDR: The discourse around AI in edtech is almost entirely about generation (e.g. what an AI Tutor says back), but IMO the real unlock is in input modalities (what and how students can now verifiably demonstrate understanding). Legacy ed tech has been fundamentally limited by point-and-click interfaces. If a student clicks around enough, they advance. Multiple choice. Drag and drop. Match the picture to the word. These interfaces can't actually verify that a student used the target skill. They can only verify that a student selected the right answer. Putting aside market/incentives misalignment, technically speaking this is why @C_Hendrick 's 1st invariant, "the target skill must be the only path to success," has been so hard to honor in software. When your only input is a click, there are always alternative paths to the right answer that bypass the skill you're trying to teach. This is why I really didn't enjoy putting my students on most edtech programs. Enough clicks = "progress" in the program that was not reproducible in any meaningful sense offline. The more interesting question is what multimodal AI implies about input. About how students can now demonstrate understanding in ways that weren't possible with point-and-click. Speech recognition means a student can read aloud and we can assess whether they actually decoded the word. Did they attend to the graphemes? Produce the phonemes? Or just guess? You can't game oral reading the way you can game multiple choice. A student can show their work on a math problem with a stylus or touchscreen that can be evaluated by multimodal AI. We can see their process, not just their final answer. How did they set up the equation? Where did their reasoning break down? This shift from clicking answers to demonstrating understanding through richer modalities is what finally makes it possible to build software that honors instructional invariants rather than routing around them. We've been building this way at @ProjectReadAI. Our AI phonics tutor requires students to read out loud. We assess down to the grapheme-phoneme correspondence. Students don't advance in the UFLI scope and sequence until they've verifiably demonstrated understanding of the focus skill. There's literally no way to click through. @burnsmk1 has some forthcoming early research on its efficacy, which is very promising, particularly for students starting below 50th pct of fluency. And @HollyLanePhD has recently provided some great feedback and ideas for how to extend this to encoding and letter formation. I genuinely think we're in the early innings of how multimodal AI can support the science of learning (and reading)

@BeCre8ive50 @HollyLanePhD @vivramak @C_Hendrick We appreciate you very much and are honored our program has been helpful.

@HollyLanePhD @vivramak @C_Hendrick What you are saying about Viv, and your experiences with him, does not surprise me in the least bit. His knowledge and skills and passion for literacy is evident in what he does. Big fan of @ProjectReadAI and @vivramak

Breaking news: Project Read has landed a major free agent signing. MaryKate Lyons brings elite elementary classroom experience and district-level support skills to our Customer Success position group. Sources confirm Lyons has stepped seamlessly into supporting schools.

@BhosalePratim this is cool! you might want to checkout @ProjectReadAI ,especially the AI Tutor.

We chat with @vivramak of @ProjectReadAI about the power—and limits—of AI in literacy instruction. From decodables to equity, this one’s packed with insights. 🎧 Listen: bit.ly/4iFcla1 📺 Watch: bit.ly/48h4FFN #AIinEducation #ScienceOfReading #EdTech

Choosing to partner with @vivramak and @ProjectReadAI was one of the best decisions we've made. They are in it for all the right reasons, which means the world to us at @UFLiteracy!

This linkedin post from a principal whose school uses @ProjectReadAI tutor made my day! When we started Project Read AI in early 2023, we had two theses: 1) thoughtfully-applied AI could level up evidence-based literacy instruction for students and teachers 2) our tiny team (2-3 folks) could lean into internal efficiencies from AI to stay tiny and reduce overhead. we could then offer free and/or extremely affordable plans for schools at scale. Besides the possibilities of what we could build together, this made partnering with @HollyLanePhD @UFLiteracy a natural fit Still a long way to go, but we've finally started growing the team and I'm really excited about what's to come

Thank you for sharing @stacy_darrow Yes, our AI Tutor that listens to students read out loud and provides real-time phonics instruction, is free for the remainder of this school year!

To wrap up our literacy block, we tried @ProjectReadAI ‘s new language game generator! Today we used language bingo as a fun connection activity - while also aligning perfectly with our @UFLiteracy lesson!

Recently got a note that moved me from a teacher with a high % of ELLs. She is using our AI tutor in her classroom. Resharing with her permission. She has now rostered parents and guardians of her students who want to learn to read English as well!!