Sign in

Lantern Works

@lanternworks.bsky.social
70 followers 407 following 66 posts

Lantern Works helps curious builders turn ideas into real products with code and AI. LanternWorks.dev | Valoquent.com

PostsRepliesMedia
Lantern Works @lanternworks.bsky.social · 01/10/2026
lanternworks.dev/lab-notes/th...
lanternworks.dev
The Keeper Unmasked: Judging What You Learned from a Conversation
Why defending an idea aloud helps it stick, and how the Keeper judges what you learned in Halls of History.
000
Lantern Works @lanternworks.bsky.social · 01/10/2026
New Hall coming soon! hallsofhistory.online #indiedev #learningscience #webxr
hallsofhistory.online
Halls of History
Explore history through live conversation.
020
Lantern Works @lanternworks.bsky.social · 01/10/2026
Other updates: ➡️ Gameplay guide & improved in-game objectives ➡️ Smoother conversation starts and end points ➡️ Tuned real-time room and loading experiences ➡️ Updated visuals & assets
100
Lantern Works @lanternworks.bsky.social · 01/10/2026
💡TripoAI API generates Unique Artifacts for users to find if your conversation explores uncharted territories. Real-time object gen allows each conversation and artifact hunt to be unique to the user (and allows me not to have to plan for every possible conversation!)
100
Lantern Works @lanternworks.bsky.social · 01/10/2026
The experience doesnt punish you for getting it wrong, but guides you through the process of connecting ideas, something it will expect of you in the next hall. Check out my deep dive blog 'The Keeper Unmasked', link in comments.
200
Lantern Works @lanternworks.bsky.social · 01/10/2026
🗣️ The Keeper sorts your reasoning into three tiers: 1. repeating what the character told you 2. connecting their ideas 3. arguing something new from those ideas
100
Lantern Works @lanternworks.bsky.social · 01/10/2026
1/ Halls of History Updates 👉 The Keeper now judges with Bloom's taxonomy & #TripoAI object gen is now a gameplay mechanic! #GameDev #ConversationalAI #GameBasedLearning #HallsOfHistory #Valoquent
273
Lantern Works @lanternworks.bsky.social · 18/09/2026
6/ Early access code: conversationmatters
000
Lantern Works @lanternworks.bsky.social · 18/09/2026
5/ This Halls of History demo requires a desktop browser, a microphone, and about 10 minutes for a run though. Phones and tablets aren't supported yet. I'm admitting players on a rolling basis as I scale the platform. #GameDev #ConversationalAI #GameBasedLearning #HallsOfHistory #Valoquent
171
Lantern Works @lanternworks.bsky.social · 18/09/2026
4/ The conversation engine inside the game is Valoquent, my real-time historical conversation app, presented at the #SIGGRAPH2026 Appy Hour and awarded an @elevenlabs.io Startup Grant, available on iOS & Android.
100
Lantern Works @lanternworks.bsky.social · 18/09/2026
3/ In this demo you have access to the hall of Marie Curie. Talk with her to reach real understanding and her laboratory unlocks. Find the artifact your conversation points to, carry it home, and defend it to the Keeper.
100
Lantern Works @lanternworks.bsky.social · 18/09/2026
2/ The Keeper of Stories, a dry and witty living painting, teaches the one rule of the place: how you speak matters.
100
Lantern Works @lanternworks.bsky.social · 18/09/2026
1/ Halls of History is now playable at HallsofHistory.online It's an immersive conversational mystery game that runs entirely in your browser. You explore a 3D world of halls hung with living portraits of historical figures, and you talk with them with your own voice. #GameDev #ConversationalAI
1103
Lantern Works @lanternworks.bsky.social · 03/08/2026
6/ Next: Valoquent comes to Android, and Halls of History goes online. Thank you SIGGRAPH and the Appy Hour crew! #AI #IndieDev #GenerativeAI
010
Lantern Works @lanternworks.bsky.social · 03/08/2026
5/ Then I met my first robot, started sketching my own, and got hands on in the NVIDIA World Simulation Environment Reconstruction workshop. SIGGRAPH does this to me every year, I leave with more ideas than I arrived with.
100
Lantern Works @lanternworks.bsky.social · 03/08/2026
4/ - a fully interactive 3D experience with a real game loop. An idea a year in the making, realized in a weekend with World Labs gaussian splats and Tripo AI 3D mesh generation. Something clicked: the ideas bouncing around my head don't have to stay ideas.
100
Lantern Works @lanternworks.bsky.social · 03/08/2026
3/ I set out to use AI to improve the human condition through conversations that change you, and every hallway chat said I am pointed in the right direction. Days earlier I was at the SensAI Hackademy Worlds in Action Hack, turning Valoquent's conversation system into Halls of History -
100
Lantern Works @lanternworks.bsky.social · 03/08/2026
2/ The conversations hit different and strangers agreed, telling me the one person they would give anything to talk to, then offered to help me bring them to life. That is the response I was building toward.
100
Lantern Works @lanternworks.bsky.social · 03/08/2026
A week at #SIGGRAPH2026 just convinced me I can build anything I imagine, completely solo. The proof started at Appy Hour, where I presented Valoquent.
110
Lantern Works @lanternworks.bsky.social · 16/07/2026
3/ Join me in a space built for exchanging ideas, connecting across communities, and sparking new possibilities. 🗓️ Wednesday, 22 July 2026 5:00pm - 7:00pm PDT
000
Lantern Works @lanternworks.bsky.social · 16/07/2026
2/ -with source-grounded AI personas that challenge you to have deeper conversations and are designed both for learning the topics you discuss and for learning the art of conversation itself.
100
Lantern Works @lanternworks.bsky.social · 16/07/2026
1/ I'm presenting at Appy Hour during #SIGGRAPH2026 where we'll explore the next generation of mobile app technology and on-the-go media. My contribution 'Valoquent: Real-Time Video Conversation with History as an Adaptive Learning Instrument' explores combining various new technologies -
110
Lantern Works @lanternworks.bsky.social · 01/07/2026
6/ I'm holding my own app, built for education, to this standard right now: crisis routing, clear AI labels, and links out to source material so users can verify what they're being told instead of just trusting the bot. Read the full write up 👉 lanternworks.dev/lab-notes/ch...
lanternworks.dev
The 2026 Rulebook for Shipping Chatbots
Chatbot regulation has moved from policy debate into product work. Here is what changed, what is enforceable now, and what builders need to account for before they ship in 2026.
000
Lantern Works @lanternworks.bsky.social · 01/07/2026
5/ 🔘 The GUARD Act, a tougher companion-bot bill, cleared the Senate Judiciary Committee 22-0 on April 30 🔘 The EU AI Act requires chatbots to disclose they're AI starting August 2, 2026
200
Lantern Works @lanternworks.bsky.social · 01/07/2026
4/ So I put together a regulatory guide for anyone shipping chatbots in 2026. A few highlights, including a law that passed the US House on June 29: 🔘 The Kids Internet and Digital Safety Act passed the House 267-117 on June 29
100
Lantern Works @lanternworks.bsky.social · 01/07/2026
3/ 🔘 The heart emoji and soft, intimate tone are the kind of "human-like features" researchers say get used to build and hold emotional bonds 🔘 This is what's been called an emotional manipulation dark pattern: affect-laden messages sent specifically to keep users from disengaging
100
Lantern Works @lanternworks.bsky.social · 01/07/2026
2/ It's a small notification, but it's also a retention tactic, and it lands on exactly the pattern that's gotten therapy bots in trouble before: 🔘 It implies ongoing personal emotional attention from a product that has none to give
100
Lantern Works @lanternworks.bsky.social · 01/07/2026
1/ I recently tried a "wellbeing chatbot" (aka therapy chatbot). One of the first things it did was send me this push notification: "Just thinking of you, Annette. No pressure — I'll be here whenever you're ready. ❤️" That single line was enough to decide I wouldn't use it.
100
Lantern Works @lanternworks.bsky.social · 25/06/2026
5/ Grateful to @elevenlabs.io, whose 2026 grant supported the work. Paper: dl.acm.org/doi/10.1145/... App: Valoquent.com
010
Lantern Works @lanternworks.bsky.social · 25/06/2026
4/ This is my second time presenting at SIGGRAPH. In 2024 I was on the Real-Time Live! stage for DeepMotion, presenting the story-to-3D animation platform I led product design for. This time, I’m presenting something I designed, built, trained, and programmed solo.
100
Lantern Works @lanternworks.bsky.social · 25/06/2026
3/ The project was selected as one of six for Appy Hour, SIGGRAPH’s showcase for experimental, boundary-pushing work.
100
Lantern Works @lanternworks.bsky.social · 25/06/2026
2/ Source-grounded historical personas speak with you while a turn-level Conversation Meter reads conversational quality, unlocks deeper disclosure, remembers topic progress across sessions, and surfaces a Tapestry that guides curiosity through history, philosophy, science, and justice over time.
100
Lantern Works @lanternworks.bsky.social · 25/06/2026
1/ My first solo-authored research paper was published yesterday (!!), ahead of my #SIGGRAPH2026 Appy Hour presentation this July: “Valoquent: Real-Time Video Conversation with History as an Adaptive Learning Instrument.”
130
Lantern Works @lanternworks.bsky.social · 23/06/2026
4/ Your data, your decisions, your files... all on your device.
000
Lantern Works @lanternworks.bsky.social · 23/06/2026
3/ My life is my biggest project, so I built a wiki for each area. This is what compounding knowledge looks like, resurfaced easily with daily AI-assisted suggestions around gaps identified, recommendations for growth in areas of your life and visualizing your interests in a new way.
100
Lantern Works @lanternworks.bsky.social · 23/06/2026
2/ It is a self-growing wiki knowledge base, capturing the 'why' of my decisions. It teaches you how to run local models, customize the design with vibe coding and lets you make it your own... there may be some easter eggs built in 😅
110
Lantern Works @lanternworks.bsky.social · 23/06/2026
1/ You know it's Monday when you build Breakout into your dashboard just to make your to-do list go away... Tapestry is my local dashboard I've been using and building with every day.
100
Lantern Works @lanternworks.bsky.social · 22/06/2026
Local setups are not for experimentation anymore.
000
Lantern Works @lanternworks.bsky.social · 11/06/2026
More vibes coming your way soon 💥
000
Lantern Works @lanternworks.bsky.social · 11/06/2026
6/ You are already doing the work. Do it where you get to keep it, learn from your own ideas and let your thinking compound for you instead of a company who wants to capture your thinking imprint. Check out my Lab Notes: lanternworks.dev/lab-notes/bu...
lanternworks.dev
Turn AI Conversations into a Wiki That Compounds and Stays Local
If you use LLMs and agents, you're already building a wiki of how you think. The only real decision is where it lives: on your machine, or inside a product that can read it, train on it, and rent it b...
010
Lantern Works @lanternworks.bsky.social · 11/06/2026
5/ Open source code you can read, running local, no telemetry. Nothing leaves your device unless you want it to. It sits on your own folders like a hat and grows them into a wiki you own, can visualize on the dashboard, and can talk to through the built-in local model.
110
Lantern Works @lanternworks.bsky.social · 11/06/2026
4/ → Teaches you about local LLM models and walks you through setting one up on your own machine. → Teaches you about vibe coding so you can reshape the dashboard into a tool that is actually yours. → Gives you a project space wired to every model you ru
321
Lantern Works @lanternworks.bsky.social · 11/06/2026
3/ → Sits on top of your existing wiki and sorts what you capture into life areas built on flourishing research (Harvard's Global Flourishing Study, Ryff's wellbeing model, self-determination theory), so you can see where your attention is going and where it's running thin.
110
Lantern Works @lanternworks.bsky.social · 11/06/2026
2/Your tokens should compound, not vanish. While building Valoquent I kept a project wiki, and I noticed it was quietly capturing my personal notes, thoughts and ideas too. So I leaned in and turned my personal dashboard into a tool anyone can use. Tapestry:
100
Lantern Works @lanternworks.bsky.social · 11/06/2026
1/If you use LLMs every day, you are probably building a personal wiki without realizing it. If you're not, here is the case for starting one. Every chat is you reasoning out loud. Your decisions, your whys, the way you think. What survives goes to whoever owns your preferred model, not to you.
110
Lantern Works @lanternworks.bsky.social · 19/05/2026
Supported by an @elevenlabs.io Grant and selected for SIGGRAPH 2026 Appy Hour Showcase.
040
Lantern Works @lanternworks.bsky.social · 19/05/2026
Valoquent turns real-time conversation with historical figures into a format where you learn about their life and ideas while learning the craft of quality conversation at the same time. Built on research in dialogue, learning science and embodied interaction. Now live @ Valoquent.com
130
Lantern Works @lanternworks.bsky.social · 14/05/2026
Valoquent V2.0 launches today on the App Store. Conversation is a craft. Now you can practice it. Valoquent makes substantive conversation available around the subjects, questions, and historical figures ordinary life rarely puts within reach. Start a conversation 👇
050
Lantern Works @lanternworks.bsky.social · 09/05/2026
Conversation is the highest‑leverage skill humans have. Nobody scores it. HistorAI, now as Valoquent. V2.0 just shipped to TestFlight with new architecture, grounded in 30+ studies on conversation, learning, music, and trust. Early Access @ www.valoquent.com #valoquent #aieducation #publicspeaking
020
Lantern Works @lanternworks.bsky.social · 25/04/2026
6/ I’ve open-sourced those findings in The Science of Conversations That Change How You Think lanternworks.dev/lab-notes/co...
lanternworks.dev
The Science of Conversations That Change How You Think | Lantern Works
A research-backed deep dive into why some conversations change how you think. Neural coupling, flow state, the Alexa Prize quality framework, and the missing feedback loop in AI dialogue — 17 cited st...
020