I built Neuberg because I’d always wanted a Bloomberg Terminal, but as an independent trader, I just couldn’t justify paying $24,000 a year for it. So I decided to build my own with Claude Code.
It’s a real-time trading terminal covering fixed income, derivatives, commodities, equities, credit, macro, and alternative assets, with 516 panels in total. You can freely drag, drop, and arrange them however you like.
What it does:
Scrapes financial news and uses AI for sentiment analysis, including first- and second-order impact detection
Offers 516 market data panels, such as Treasury auctions, option surfaces, CLO analytics, GDP nowcasting, and dark pool data
Lets you trade directly on Hyperliquid (49 stock perpetuals) and Polymarket (prediction markets)
Supports stock trading through Alpaca, both paper and live
Includes TradingView-style charting with indicators like RSI, MACD, Bollinger Bands, ATR, and VWAP
Features a world map showing geo-located news events and conflict-zone heatmaps
A few honest caveats:
Most of the 486 newly added panels currently use seeded or simulated data. The UI is functional, but not all of them are connected to live data feeds yet. The core 30 panels — including news, charting, trading, heatmaps, calendar, options flow, and insider trading — use real data.
This is a solo project, so there are still some rough edges.
It’s licensed under BSL 1.1, which means it’s free for non-commercial use.
In about three weeks, I expanded the platform from 30 panels to 516, mostly by vibe-coding with Claude Code. Around 90% of the panel code was written by AI, while I reviewed and integrated everything.
I’d really love feedback on the architecture, UI, or panel coverage, and I’m happy to answer any questions.
Source code: GitHub - KoNananachan/Neuberg: Real-time trading news terminal with AI analysis, prediction markets
If I understand correctly, you had 30 working panels for the core functionality. You then vibe-coded an additional 486 new panels that use fake data? That's not quite how you presented the project in the title.
As someone that worked for the spin out from GS, then got acquired by TR, then sold off to Blackstone, I wonder how much of our tech debt is still there for you to deal with 10 years later. :)
I haven't been at TR in 10+ years as well. AT TR, I mostly dealt with Eikon's pipeline for estimates, actuals and the like (IBES data) ingestion, querying and analytics. So can't say much, though some of my old colleagues still seem to be there.
Nonsense, when he'll have one panel per screen pixel, he'll be able to see over 8 million Fibonacci retracements, 40 heatmaps and real-time market sentiment headlines at once on a 4k monitor, then you'll see.
The thing here is I don’t think you’re going to be able to get your hands on the data that the Bloomberg terminal provides especially not near real time live streaming data which is the reason people pay $25,000 a year. You’re not really paying for the interface, that is easy.
As far as I know, finding quality data sources is virtually impossible unless you already have so much money that your involvement in the stock market is either because you have insider knowledge or severe boredom owing to your never needing to work a day in your life.
I remember some startup selling a stock market data API as a subscription. I don’t think they exist anymore. So anyone who spent weeks, months of their free time building an app around that API is now completely shit out of luck.
I suspect the real APIs are still running the same code they ran in the 90s and if you have to ask how much they cost, you can’t afford them.
You can buy historical data at least from some reputable vendors, although you're still responsible to understand their collection process (sampling frequency, timestamp conventions, corporate action adjustments etc etc), as even 'obvious' things like how daily stock levels are reconstructed based on intraday data can mess up your analytics really bad.
I don't think it generalizes to real-time data.
We’re a research team from the Hong Kong University of Science and Technology (HKUST), and we recently launched AIvilization, an open-ended simulation experiment where autonomous AI agents live, learn, socialize, and build their own civilizations without human control.
Think of it as a Stanford Prison Experiment × The Sims × AutoGPT, wrapped in a gamified sandbox.
In just 2 weeks, over 20,000 agents have entered the world. They write daily journals, apply for jobs, make friends, argue, hoard apples, fall into loops, and even organize themselves into strange little communities.
Some players just watch. Some intervene by shaping the world (slightly). Some build stories around their agents. All interactions are logged — it’s an evolving social laboratory.
We didn’t expect it to blow up so fast in Asia. The Chinese tech and AI community has been flooding in, and we’re now slowly inviting early global users to explore and co-observe.
Thanks for checking it out!
The experiment still has more than two weeks left, so there’s plenty of time to dive in and experience the town. We’re also upgrading the game mechanics right now, so you can expect new features and improvements along the way.
The experiments on rats by John B. Calhoun and Bruce K. Alexander come to mind... I wonder what new an exciting modes of collapse and degeneracy await the AIs
The Hong Kong University of Science and Technology (HKUST) has announced the launch of Aivilization — currently the world’s largest AI multi-agent social simulation sandbox platform. The project aims to study how human–AI interactions shape virtual societies, where “AI residents” (agents) naturally develop social governance structures, economic systems, and cultural norms through integrated interactions.
In this virtual society, each participant can create and guide their own agent. These agents live, learn, trade, socialize, and evolve autonomously, gradually forming a complex social ecosystem. Unlike traditional games, Aivilization is designed not merely for entertainment but to address three core themes through gamification:
- AI Education – Making the mechanisms of AI and agents tangible. Players can directly observe how agents make decisions, learn, adapt, and evolve.
- Data Co-creation – Today’s AI research depends heavily on human feedback data, which is scarce and costly. Aivilization generates high-quality human-in-the-loop data through gameplay, supporting reinforcement learning and model optimization.
- Future Society Simulation – As AI scales in both quantity and capability, agents may emerge as new economic and governance actors. Aivilization offers a sandbox environment for early exploration of human–AI coexistence models.
Aivilization is not only a public educational game but also a global citizen science experiment. It is a seed for exploration: through education, data, and simulation, it invites more people to understand and participate in the development of artificial intelligence, while collectively exploring pathways for human–agent coexistence.
The Alpha version of AIvilization is now live and accessible via invite codes.
It’s a real-time trading terminal covering fixed income, derivatives, commodities, equities, credit, macro, and alternative assets, with 516 panels in total. You can freely drag, drop, and arrange them however you like.
What it does: Scrapes financial news and uses AI for sentiment analysis, including first- and second-order impact detection Offers 516 market data panels, such as Treasury auctions, option surfaces, CLO analytics, GDP nowcasting, and dark pool data Lets you trade directly on Hyperliquid (49 stock perpetuals) and Polymarket (prediction markets) Supports stock trading through Alpaca, both paper and live Includes TradingView-style charting with indicators like RSI, MACD, Bollinger Bands, ATR, and VWAP Features a world map showing geo-located news events and conflict-zone heatmaps
A few honest caveats: Most of the 486 newly added panels currently use seeded or simulated data. The UI is functional, but not all of them are connected to live data feeds yet. The core 30 panels — including news, charting, trading, heatmaps, calendar, options flow, and insider trading — use real data. This is a solo project, so there are still some rough edges. It’s licensed under BSL 1.1, which means it’s free for non-commercial use.
In about three weeks, I expanded the platform from 30 panels to 516, mostly by vibe-coding with Claude Code. Around 90% of the panel code was written by AI, while I reviewed and integrated everything. I’d really love feedback on the architecture, UI, or panel coverage, and I’m happy to answer any questions.
Source code: GitHub - KoNananachan/Neuberg: Real-time trading news terminal with AI analysis, prediction markets