367 karma · joined July 28, 2012
The engine is JEV, TypeSafe's "System One" model, served through OpenRouter. It never writes text: you hand it a state and a typed question, it hands back a probability for every option. Every legal move goes over as the options of one question, so an illegal move isn't blocked - it's impossible. The bars on the board are the probabilities exactly as returned; on easy the top move just gets played.
The part I find most interesting is the "Is it calibrated?" panel: every move JEV has ever played, bucketed by the confidence it reported at the time, scored by whether the move held up - i.e. it did not give material away, judged by the same one-ply check the model was shown. A narrow test, but the same one every time, and it updates live (422 moves scored so far).
Details:
- Easy is one answer, ~400 ms. Hard also plays your reply in its head and ranks that too. - Material counting happens in code and is fed back in as state - JEV is weak at arithmetic and I don't pretend otherwise. - It remembers openings it has played and how those games ended, fed back in as state. - One shared board: everyone plays the same side against it, first click wins the turn. "Start a private one" gives you your own game.
Score so far: JEV 15.5, humans 2.5 (elo ~1317, easy mode). Total API spend to date: $0.07. Free to play, no account needed. Your move.
https://news.ycombinator.com/item?id=47533035
:)
> How do I see the stats of a city? Click the city tile.
> When I use the "mine" tool, what happens? The nearest dwarf attempts to so the task? Dwarf live in mines, this is city extension tool basically.
But I push back on "the entire premise is wrong."
The interesting part isn't "AI can execute pseudocode"—nobody debates that. The point is the artifact: the .md output matters, not the runtime. A codebase where every function is readable English changes who can participate in a pull request, audit logic, or catch wrong assumptions. "Multiply by 9/5" vs "multiply by 1.8" is an editorial conversation, not a code review.
It's a proof of concept to show the artifact is executable, not a production proposal. It's slow (today), expensive, and non-deterministic - I said so in the post. The question is whether the intermediate representation (English) has value beyond performance? I believe that the loop is shortened here: there're no in-between element intent→weird non-human language→result, it becomes intent→result. No NEED to create synthetic procedures, explaining how the code works in plain language should give us the output.
An old person enters a bank, asks to open an account, speaks plainly in his native language, the teller clicks buttons, and the account is created. From the subjective perspective, there's no in-between interface: the old person had a though, than it got realized.
Consider this scenario: The company hires remotely and receives 1,000+ resumes. Most are spam and irrelevant. It would take the team days to read and process them.
A few days ago, I developed an AI-driven tool to enhance my hiring process. It features:
- PDF text extraction (OCR, Computer Vision) - NLP-based resume and job description matching - Match percentage and compatibility analysis - Candidate feedback generation - Red flag detection and portfolio analysis (I check for 404 errors) - Candidate ranking with emojis and titles
Leveraging OpenAI's GPT and Anthropic's Claude (added OpenAI support recently), it automates screening and delivers actionable insights. After evaluating 100 resumes, I've identified a strong candidate.
I get the irony, this might seem a bit unfair, but it’s also a showcase of how these smart HR systems operate—black boxes by nature. I’m looking for feedback on how well it matches candidates (and weeds out spam) and would love any suggestions for additional features. What are your thoughts on AI’s role in recruitment?
Development journey: https://x.com/stas_kulesh/status/1838524611513708940
Questions or insights welcome!
Key features: - Extracts text from PDFs (OCR, Computer Vision) - NLP-based resume-job description matching - Provides match percentage and compatibility breakdown - Offers candidate feedback - Checks for red flags and analyzes portfolios - Ranks candidates with emojis and titles
Using OpenAI's GPT and Anthropic's Claude, it streamlines recruitment by automating screening and providing insights. I've screened 100 resumes and found a promising candidate.
I'm seeking feedback on the matching accuracy and ideas for new features. How do you think AI will impact hiring?
Thread on dev process https://x.com/stas_kulesh/status/1838524611513708940
Happy to answer questions and hear your thoughts!
Think of it as a tongue-in-cheek tribute to the complexities and idiosyncrasies of building a groundbreaking AI company, with a hint of OpenAI flavor.
The game puts you in the shoes of a startup CEO on a quest to achieve Superintelligence, challenging players to navigate through a minefield of industry-specific issues like ethical dilemmas, data biases, and the ever-present specter of regulatory compliance. It's a bit like playing Dungeons & Dragons, only your quests involve balancing AI power growth with cash flow to keep your talent from jumping ship.
I created this as a playful commentary on the tech industry's quirks and our collective fascination with AI's potential and pitfalls. It's a strategic game with a dose of humor and a reflection on the startup grind.
Why did I make it? Partly as a creative outlet, partly as an homage to the tech culture that many of us here are deeply immersed in. It's for anyone who's ever chuckled at the absurdity of tech buzzwords or rolled their eyes at the latest AI hype.
Would love to get the community's feedback, and if anyone has stories or strategies from playing the game, I'm all about swapping tales from the trenches.
You can check it out here. I look forward to hearing your thoughts and experiences.
On a positive side, I believe it's a matter of time. In a couple of months it'll be cheaper or free for all.
The custom GPT resume writer will convert your existing PDF or plain text content into an organized, well-formatted, rewritten, and ATS-optimized version of your resume. It really fulfills many parameters with its enhancements. It provides actionable recommendations and explains itself in detail.
The enhanced resume, or CV if you're in Europe, can be used for work, job searches, and more. Most importantly, I'm trying to help people tailor their resumes to specific vacancies. You can share the link in a chat, or you can copy and paste the requirements from the new job position description, and the Resume Writer will produce a well-formatted, clean markdown resume optimized for that particular position.
And, I guess, this is the main thing that happens in ChargeGPT (if you have access to ChatGPT Plus, unfortunately), the GPT itself comes for free.
You can always take that enhanced resume out of the chat and do more with it using resume-building software-as-a-service tools. The Resume Writer GPT is powered by the latest GPT technology and simply does the job.
Enjoy.
Title: Show HN: Bringing Retro Gaming to ChatGPT with a Might & Magic-Inspired Text Adventure
I've been captivated by the idea of creating a text-based game within ChatGPT, turning it into a nightly hobby. Every evening, I dive into this textual realm, inputting commands and marveling at the unfolding adventure. This project has evolved into a nostalgic ode to the retro games of the '90s, particularly the Might & Magic series (IV-VII) that I enjoyed during my high school years on a local PC with friends.
This game, leveraging the latest ChatGPT updates, aims to resurrect the classic gaming experience, where you lead a group of characters through various locales, battling monsters along the way. I'm excited to demonstrate how ChatGPT's creativity can not only match but enhance my own. I hope you find it as engaging and fun as I do.
I welcome any questions or suggestions for improvement. Your feedback is invaluable as I continue to refine this interactive experience.
Latest changes and the process described here: https://x.com/stas_kulesh/status/1725278103847489642?s=20