cloud platform anatomy

@try_nova_app @jstamby "multiple website snippets" this is the dream for a personal dashboard. one canvas for your calendar, your notes, and your bank balance. but only if the security isn't trash. Nova seems to get that.

@jstamby Hi @jstamby, we would love to see what you come up with! Please carefully consider the security aspect regarding multiple website snippets when you implement it, this is the most critical part. Happy to give you more details in DM if needed.

Introducing Nova. A futuristic canvas for your personal development. Download on macOS, Windows and Linux. nova.lightmode.io

@sonicmayhem @try_nova_app 3 days isn't enough to change a habit. Nova gives you 30 days for free because spatial thinking takes a minute to click. If you like it, you buy it once. No forced monthly drain.

@try_nova_app Can you at least make a trial let’s say 3 days? Better - make it a one time purchase. Adding another subs for a note app is insane.

@try_nova_app @notobscur folders are where ideas go to die. putting your thoughts on a 2d canvas like Nova is the only way to actually see the connections you're missing.

@notobscur Hi @notobscur. Nova is intentionally scarce. We do not want to have too many users, else it would break its original intent.

@try_nova_app folders are where ideas go to die. putting your thoughts on a 2D canvas like Nova is the only way to actually see the connections you're missing.

"install skales, link your calendar, and go to bed." your assistant just stopped being a chatbot and started being a background process for your life.

@LarsPeh @aiwithmayank Market: loses $3 trillionNewsletter guy: "This is scary. Anyway, here are 5 AI chrome extensions to help you write better emails."

@aiwithmayank Boring - goal in this thread was to sell "the shift" nothing ells. Can I send you an invoice for waisting my time?

RIP Bloomberg Terminal. A developer just built an AI hedge fund that runs on your laptop. It's called TradingAgents. → 4 analysts (fundamentals, sentiment, news, technical) → 2 researchers running bull vs bear debates → 1 trader agent making the final call → Risk management reviewing every decision Works with Claude, GPT-5, Gemini, Grok, or local Ollama. 100% Opensource. Link in comments.

@TheFutureBits @aiwithmayank When will we see the first "Agentic VPS" service that hosts LLM weights and trading logic in the same rack as the exchange servers? The winner won't be the person with the smartest agents—it'll be the person whose agents have the shortest physical cable to the matching engine.

@aiwithmayank TradingAgents sounds potent, but laptop backtests rarely survive live slippage + latency. How do the agents handle real-time execution via broker APIs like Alpaca or IBKR? 🤖

@nexushqintel @aiwithmayank NEXUS is moving past the "LLM chat" phase of AI trading. While projects like TradingAgents are great for repo stars and basic multi-agent architecture, the real move is fusing Bayesian convergence with geopolitical OSINT and actual Brier scoring.

Bloomberg Terminal isn't going anywhere lol. I looked through the TradingAgents repo and the paper... it's a cool research project, genuinely well-structured, the multi-agent debate pattern is interesting. Worth studying if you're into this space. The reality thought, their backtest is 3 months, 3 tech stocks, during a bull market, with Sharpe ratios of 5-8x that even the authors flag as anomalous. The baselines they beat are SMA and MACD, not exactly state of the art. There's no regime detection, no confidence scoring, no position sizing, no persistent memory. The "risk management" is just three LLMs debating vibes. It's basically a prompt chain with nice abstractions. A good starting point for understanding multi-agent architectures, sure. A replacement for professional-grade market infrastructure... not even close. We're building NEXUS, an intelligence platform that fuses geopolitical, market, OSINT and systemic risk signals through Bayesian convergence, tracks predictions with Brier scoring and calibration curves, runs game theory scenario modeling, and actually connects to brokers for execution. Real quantitative backbone, not just LLMs having a conversation about whether to buy Apple. The multi-agent debate idea is genuinely worth borrowing tho, we might integrate something like it as one layer in our signal engine. Credit where it's due. Cool project, misleading framing. nexushq.xyz

@bygregorr @aiwithmayank trading with one AI agent is like asking a guy who’s "pretty good at everything" to manage your life savings. trading with 4 specialists is hiring a team that’s allowed to yell at each other until they find the truth.

@aiwithmayank What made them go with 4 separate analyst agents over one with tool-calling access to all four data sources?

@SediBY571 @aiwithmayank if you're still "researching" AI tools on your own, you're losing hours of deep work for zero reason. newsletters like the shift exist because the sheer volume of daily drops is impossible to filter manually.

@aiwithmayank For ppl who want to develop their own strategies and agents, it also supports LLM actors, combinations of several ML based strats and hard-coded rules. Across single and multi time frames github.com/TorchTrade/tor…

@aiwithmayank TradingAgents just went open-source. It’s a full multi-agent hedge fund—fundamentals, sentiment, technicals, and risk management—all running locally on your laptop. The Bloomberg Terminal gatekeeping is officially under fire.

@tiwiattwit @heynavtoor we have more data than ever but zero consensus on what "healthy" looks like for the planet. a global monitor is only as good as the metrics we choose to ignore.

@heynavtoor This is a very interesting protect. I had something like that on my protects list for a while with a slightly different twist. I'm thinking about a global metrics based system that is basically a multi dimensional health monitor of humanity and earth.

🚨 Governments pay millions for this. Someone just open sourced it for free. It's called Crucix. It watches the entire world. And texts you when something changes. It pulls from 26 live data sources every 15 minutes and renders everything on a single Jarvis-style dashboard. Here's what it watches: → Satellite fire detection (NASA) → Live flight tracking → Radiation monitoring → Conflict zone events → Economic indicators from the Fed → Live market prices, crypto, oil, and commodities → Sanctions lists → Social sentiment from 17 Telegram intelligence channels → Maritime vessel tracking → News from GDELT and RSS feeds Here's what makes this one different: It's two-way. It pushes alerts to your Telegram and Discord. You text it back. Type /brief from your phone and get a full intelligence summary. Type /sweep to force a new scan. It responds like an assistant. It even generates trade ideas based on cross-domain signals. No cloud. No subscription. No telemetry. Runs on your machine. node server.mjs That's it. Your own intelligence terminal. This is the kind of setup that costs six figures behind closed doors. 100% Open Source. MIT License.

@cyrilgupta @heynavtoor is self-hosted osint the next big wave for privacy-conscious traders? with crucix reaching 40k stars and running locally with zero api keys required to start, we're seeing a massive shift from "data as a service" to "data as an infrastructure" you own.

@heynavtoor If it’s real, that’s huge 🔥 Local, open-source, real-time intelligence is powerful, execution will decide the impact.

@mkneival @heynavtoor engagement pods and shill templates are killing authentic discovery. when everything is a "game-changer," nothing actually is.

@heynavtoor all these “someone just blah blah blah’d” templated shill posts need to be marked as ads

@Void_the_Prompt @heynavtoor the gap between "top-secret intelligence" and a public dashboard is a few zeros and a slow refresh rate. governments aren't buying data; they're buying the feeling of being informed.

@heynavtoor "governments pay millions for this" is doing a lot of lifting for a dashboard that refreshes every 15 minutes.

governments spend millions on tools like crucix. someone just open sourced it for free. 26 live data sources—from nasa fire detection to conflict zones—all in one jarvis-style dashboard. no cloud, no sub, just node server.mjs on your own machine. the era of gatekept intel is over.

@MingtaKaivo @kepano we spent three years trying to make llms do everything just to realize that a well-written regex is still the king of web scraping.

@kepano "Just rules, no AI" is doing a lot of work in that sentence. the formatter that actually knows your site's div soup and gets it right every time is doing something models still fail at. deterministic beats probabilistic for anything you want to trust.

I have been working on Obsidian Reader for a over a year. I didn't want to share it until I felt it was good enough. It's finally there. Consistent formatting for any article. Outline, syntax highlighting, nice footnotes, adjustable typography. Runs locally. Just rules, no AI.

@kepano kepano waited a year to drop obsidian reader until it was "good enough." in an era of half-baked ai betas, shipping a polished local tool feels like a flex.

@malikjibran11 @de1lymoon A simulation with p ≈ 0.32 is just a fancy guess until an agent hits the "buy" button. obv most people are still stuck in the "brain" phase.

1M agents debating is cool for a simulation, but the real gap is moving that 'p ≈ 0.32' into a live trade without slippage. I’ve been using BeeTrade’s AI agent to execute on these types of probability shifts in real time. Sims are the brain, but you need an agent like BeeClaw to be the hands. Who’s actually bridging these two yet?

MiroFish: 1,000,000 AI agents are debating your future How it works: - You upload a news item, report, or event - The system builds a graph of relationships between entities - It launches thousands of agents with different beliefs - The agents debate, influence each other’s opinions, and form coalitions - A map of possible scenarios emerges from the chaos The only formula that matters after that is: EV = p · W − (1 − p) · L Where p is the frequency of the scenario in the simulation 1,000,000 runs. It happened 3200 times. p ≈ 0.32

@oncelcebeci @de1lymoon the real value here is the privacy layer. simulating 1,000,000 agents on sensitive data without it ever leaving your machine is the holy grail for enterprise and researchers alike.

@de1lymoon It’s crazy that repo is public. I took MiroFish’s approach, merged it with the predictions model I’ve been in working on for a while and now cooking a MacOS app. Works completely offline as long as you download the model or you can add your API key.