bturtel
A small economic forecaster trained from raw Fed PDFs beat GPT-5
Show HN: Open-source LLM and dataset for sports forecasting (Pro Golf)
https://huggingface.co/LightningRodLabs/Golf-Forecaster
Hey HN, I fine-tuned a small open-source model on golf forecasting and it beats GPT-5 at predicting golf outcomes. The same approach can be used to build a specialized model in any domain, you just need to update a few search queries.We fine-tuned gpt-oss-120b with LoRA on 3,178 golf forecasting questions, using GRPO with Brier score as the reward.Our model outperformed GPT-5 on Brier Skill...
Re:Show HN: Trained an LLM to predict "What will Trump do?"
on Friday February 20, 2026 @12:56PM (#47091384)
Great question! It's probabilistic so not really "right vs wrong" on any single question, but who better estimated the likelihood.
One big difference shows up when there's no useful context - we ran the same eval WITHOUT including any useful up-to-date context with questions. In this case, GPT-5 stays overconfident and its BSS drops to -11.3% (vs -4.3% ours) - worse than just guessing the base rate.
So one advantage of the RL training is just learning to know what you don't know, and identify when there's real signal.
Show HN: Trained an LLM to predict "What will Trump do?"
https://huggingface.co/LightningRodLabs/Trump-Forecaster
Hey HN! I RL-tuned an open-source LLM (gpt-oss-120b — 120B MoE, but only 5.1B active params) to predict "What will Trump do?" in any situation, trained on nothing but public news collected automatically from search queries. The trained model beats GPT-5, and both dataset and trained model are open sourced.Data generation: Generated 2,108 binary forecasting questions from just a search query and...
What can't be automated? The Last Human Bottleneck
Future-as-Label: Scalable Supervision from Real-World Outcomes
TMLR: Outcome-Based Reinforcement Learning to Predict the Future
Natural Selection Is Already Shaping AI
Flooding the AI Frontier
Foresight-32B Beats Frontier LLMs on Live Polymarket Predictions
Why Apple Will (Eventually) Win AI
Outcome-Based Reinforcement Learning to Predict the Future
What remains scarce post-AGI?
Why humans are still much better than AI at forecasting the future
Re:LLMs Can Teach Themselves to Better Predict the Future
on Tuesday February 11, 2025 @06:58PM (#43019956)
Great question!
The key advantage of self-play is that we don't actually have labels for the "right" probability to assign any given question, only binary outcomes - each event either happened (1.0) or did not happen (0.0).
Our thinking was that by generating multiple predictions and ranking them by proximity to the ground truth, self-play incentivizes each agent to produce more finely calibrated probabilities - or else the other agent might come just slightly closer to the actual outcome.
Re:LLMs Can Teach Themselves to Better Predict the Future
on Tuesday February 11, 2025 @02:59PM (#43017601)
We're working on a follow up paper now to show similar results with larger models!
Re:LLMs Can Teach Themselves to Better Predict the Future
on Tuesday February 11, 2025 @01:41PM (#43016528)
Great read! Thanks for sharing.
LLMs Can Teach Themselves to Better Predict the Future
Show HN: Sculptor – Python library for LLM structured data extraction (MIT)
https://github.com/lightning-rod-labs/sculptor
I built Sculptor after repeatedly seeing founders try to hire data scientists for a task that ultimately boiled down to extracting structured data from unstructured text (customer records, social posts, websites, etc) using an LLM API. We ended up reinventing this pattern internally at least three times in the past year, so I published Sculptor as a streamlined, open-source solution:- Simple...
Re:Show HN: Log each meal with a photo to track your diet and monitor ...
on Monday January 13, 2025 @11:09AM (#42684985)
This could be huge. IIUC this is basically an AI-enabled version of MyFitnessPal, which has like 200M+ users, but with a massively streamlined user experience. Great idea.
Re:Show HN: I made an app to challenge anxious thoughts with AI
on Monday January 13, 2025 @10:53AM (#42684783)
This looks really cool - the UI in particular feels really approachable and polished. I like how you detect and call out "What you're doing wrong" to help build awareness of unhelpful thought patterns. Upvoted!
I recently launched something pretty similar (www.pensiveapp.com) but we took a very different approach (and I think the space is huge). Interesting to see how differently you approached the problem.
Re:Show HN: Argument Encyclopedia: User-Posted Claims and Rebuttals
on Monday January 13, 2025 @10:48AM (#42684728)
This is really cool - I think its really helpful in difficult conversations when you can encourage people to choose a single branch / claim of the argument and stick to resolving that before confounding by mixing in other claims.
Personally, I'd love to be able to see a visual branching of different arguments / counterarguments, with some visual indicator for how much support each has.
Re:Show HN: Demo of my web game about social persuasion
on Sunday December 22, 2024 @04:45PM (#42489397)
Yea, I think so - I could imagine this being really streamlined by just dropped me immediately into a conversation, with maybe the goal just written on a screen somewhere - no setup, no storyline, etc. I guess it just depends if most of your users are there for a gameplay experience vs a "practice" experience.
Re:Show HN: Demo of my web game about social persuasion
on Saturday December 21, 2024 @05:51PM (#42482976)
I think this has a TON of potential. Situations like these are very non-obvious and anxiety-inducing for lots of people, so if you can make this a way for people to gain proficiency and confidence at navigating tricky social interactions, it could be a very powerful value prop.
My only feedback would be that it took too long to get into the first challenge - lots of instructions / introduction / scene setting.
Well done!
Re:Show HN: Pensive – AI mental health coaching backed by science
on Monday December 09, 2024 @01:18PM (#42368824)
Thanks! Great questions.
We just launched - anecdotally we've had really great feedback from early users, but we're working with PhDs in the field to design an external validation study while tracking user-reported outcomes. We've also heard from a few providers that have started recommending Pensive to their clients for on-demand support between sessions.
We see Pensive as complementary to therapy for some users and a standalone tool for others. Many people who don't need therapy can still benefit significantly from consistent, evidence-based practice. The key is making these proven techniques more accessible. Like physical health, most people don’t need a medical intervention - they need to exercise, but sorting through the research and implementing effective practice is a big barrier.
Our focus is on delivering established practices in a convenient format with just a few minutes of guided conversation a day.
Show HN: Pensive – AI mental health coaching backed by science
https://www.pensiveapp.com/web-demos
Hey HN! We turned proven practices (from CBT, DBT, Stoicism, and more) into guided 5-minute conversations with a personalized AI coach.Traditional care is broken: expensive ($200/hr), inconvenient, and often unavailable when needed most. Whether you're in therapy or not, most of us can significantly enhance well-being with evidence-based practice.Pensive leverages research-backed techniques and...
Re:Sunlight is more effective than censorship
on Wednesday November 20, 2024 @09:42AM (#42194289)
I'm not sure what you're criticism here is - the article never uses the term "fake news", and it gives specific examples of factual inaccuracies promoted by the government.
I fully agree that the cryptocurrency space if plagued with scams and theft, but I don't think that should prevent us from seeing the benefits of open & verifiable data that it has the potential to deliver.
Sunlight is more effective than censorship
Re:Show HN: Model2vec – Lightning-fast Static Embeddings for RAG/Seman...
on Tuesday November 19, 2024 @06:58PM (#42189389)
This seems awesome for enabling RAG queries for on-device LLMs.