Almusty Labs (@AlmustyLabs) header
Almusty Labs (@AlmustyLabs) profile picture

AI Agents • Automation • Open Source AI Tools, insights & experiments. ⚡ DM for collaborations.

Chicago, IL
Joined February 2020
Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)Photo by Almusty Labs (@AlmustyLabs)
Almusty Labs retweeted
Here's a post on what Congress / the White House should do to proactively head off a huge rise in cyber damages due to AI; this is a summary of the congressional briefing I gave on the hill yesterday with @ReadyStateLabs and Senator Chuck Schumer's office. joshuasaxe181906.substack.co…
2
20
873
Almusty Labs retweeted
The whole hobby, honestly
1
23
Almusty Labs retweeted
The real work in trading automation happens before the first trade. Clear rules, proper limits and knowing what the bot should do when things don't go as planned.
Most people lose money automating @Polymarket strategies because they skip the boring parts. Here’s the exact order we force every system through before it ever touches real money: ▫️WRITE THE RULE One sentence a stranger could execute. Example: “Buy YES when the ask is under 40¢ and the model says ≥50%.” ▫️LIST THE INPUTS Price, model probability, balance, end date. Where does each one actually come from? ▫️DECIDE WHAT HAPPENS WHEN AN INPUT IS MISSING Skip and log the reason. Never guess. ▫️SET THE LIMITS Per-trade size. Per-market cap. Total exposure. Max slippage. ▫️HANDLE THE ORDER LIFECYCLE Placed → partly filled → cancelled → rejected → unknown. Every state needs a next action. ▫️ DRY-RUN FIRST Log “would place.” Post nothing. Compare it to what you would’ve done by hand. Every mismatch is a bug. Only after all six do we go live, small. If your strategy is already written as rules, DM us “BUILD” + the rules + what’s still manual. Most bots fail because the rules were never finished. We don’t ship unfinished systems. 🦁
1
2
27
Almusty Labs retweeted
This was a great success So I've also added a report endpoint to report AI reply bots yourself blockedforaireply.levels.io/ Feel free to add them to your Chrome extensions etc. My AI reviews them all, and then I review them again to see if their other replies are also AI and then if confirmed add them to the block list
It took a while but as promised here it is... "THE 𝕏 SPAMMERS LIST" 🛡️ Blocked for AI reply blockedforaireply.levels.io 3,728 𝕏 accounts that continuously spam automated AI replies to artificially farm engagement and ruin 𝕏's social network - 498 already suspended by 𝕏 - continuously updated ✅ Evidence for every blocked account ✅ One-click block all button ✅ Continuously updated JSON API ✅ Text-only username list Based on @ryanvogel's original Clankwall and @dillon_mulroy's X list I hope everyone uses this and keeps using it, so my account becomes a honeypot for AI replies, once I block them they're blocked by the entire network and they stop being able to reply to anyone + It'd be cool to add more "scouts" like @ThePrimeagen and @JackEllis and sync their block list continuously too cause they and many others have the same issue I've also submitted this list to @X's CEO @petersellis to train their spam models on Again it'll be continously updated! Enjoy a spam free 𝕏 experience! 😊🙏
22
6
5
127
16,014
AI can organise market research summarise reports, pull historical data, and flag events worth looking at. It still can’t assess the risk for you. An organised summary isn’t a prediction, and the decision still sits with the person reading it. Have you used an AI research tool that made the information cleaner but still left the hard judgment calls to you?
5
62
Almusty Labs retweeted
Asking AI to confirm what you already believe usually just gives you a cleaner version of the same view. It’s more useful to ask it to find evidence that contradicts your assumption. The value is in the pushback, not the agreement. Have you deliberately asked an AI to challenge one of your assumptions and actually changed your mind?
6
12
27
382
Almusty Labs retweeted
An AI market summary can sound current even when the news is weeks old. If it doesn’t show the original date and a source link, it’s easy to treat last month’s announcement as something that just happened. Freshness matters as much as the wording. Have you seen an AI summary present old news as if it were recent?
5
2
14
374
Almusty Labs retweeted
AI can attach a source to a claim and still misrepresent what that source actually says. The useful habit is asking it to point to the specific part, then opening the source yourself. A linked source doesn’t automatically mean the claim is accurate. Have you checked a source an AI cited and found it didn’t fully support the statement?
17
2
22
416
Asking AI to confirm what you already believe usually just gives you a cleaner version of the same view. It’s more useful to ask it to find evidence that contradicts your assumption. The value is in the pushback, not the agreement. Have you deliberately asked an AI to challenge one of your assumptions and actually changed your mind?
6
12
27
382
AI can attach a source to a claim and still misrepresent what that source actually says. The useful habit is asking it to point to the specific part, then opening the source yourself. A linked source doesn’t automatically mean the claim is accurate. Have you checked a source an AI cited and found it didn’t fully support the statement?
17
2
22
416
An AI market summary can sound current even when the news is weeks old. If it doesn’t show the original date and a source link, it’s easy to treat last month’s announcement as something that just happened. Freshness matters as much as the wording. Have you seen an AI summary present old news as if it were recent?
5
2
14
374
Almusty Labs retweeted
Most Web3 project announcements sound stronger than the actual product. Before treating the claims seriously I check the docs, the repository (if it exists), and whether the network or product is actually doing what the marketing says. A popular post isn’t evidence. Have you looked into a project that sounded impressive in announcements but looked weaker once you checked the code or docs?
15
5
22
417
Almusty Labs retweeted
AI agents become more interesting when they move beyond answering questions and actually complete tasks. AIApply is a good example of that direction. It finds relevant job opportunities, prepares applications with your resume and a role-specific cover letter, and lets you decide whether to review each one or automate the process. Less repetitive work. More control over how you apply. This is the kind of practical AI automation worth paying attention to.
We just killed job hunting. Introducing AiApply: a career agent, in your pocket. Tell it what you want. It scans millions of jobs, skips 99.9% of them, and helps you apply to the few that perfectly fit, with a resume and cover letter written for each. • No Spray and Pray. Our matching learns from millions of applications and interview outcomes. • Accurate. Every application uses your real skills and experience. Nothing invented. • Swipe to Apply. Every swipe teaches your agent. • Full Control. Review every application, or put the strongest matches on autopilot. • Already Applied. Other apps alert you when the right job goes live. AiApply has already applied. Link below.
47
3
1
55
565
Almusty Labs retweeted
Most AI projects look better on social media than they do in the actual documentation or code. I usually check four places before taking the claims seriously: the official site, the docs, the GitHub repo, and the technical paper or model card. Social posts are useful for discovery. They’re weak for verification. What’s something you believed about an AI project from social media that changed after you checked the source?
7
2
25
420
Most Web3 project announcements sound stronger than the actual product. Before treating the claims seriously I check the docs, the repository (if it exists), and whether the network or product is actually doing what the marketing says. A popular post isn’t evidence. Have you looked into a project that sounded impressive in announcements but looked weaker once you checked the code or docs?
15
5
22
417
Almusty Labs retweeted
AI can summarise research quickly, but a fast summary can still mix verified facts with claims that aren’t fully supported. The useful part is being able to tell which points are backed by the sources and which ones still need checking. Have you caught an AI research summary that sounded solid but included something the original sources didn’t actually say?
10
4
24
562
Most AI projects look better on social media than they do in the actual documentation or code. I usually check four places before taking the claims seriously: the official site, the docs, the GitHub repo, and the technical paper or model card. Social posts are useful for discovery. They’re weak for verification. What’s something you believed about an AI project from social media that changed after you checked the source?
7
2
25
420
Almusty Labs retweeted
A small update on my ZOQO Build-Off journey. I've started experimenting with automated BTC 5-minute Up/Down trading using ZOQO. My first strategy was Target Detector. It monitors BTC price movements and automatically enters trades when certain conditions are met. But here's something interesting I discovered: My automation recorded an 80% win rate, yet its P&L was still negative (-$0.79). That was a reality check. A few lessons so far: → A high win rate doesn't guarantee profit.
→ Entry prices matter.
→ Liquidity affects execution.
→ Risk management is essential.
→ Backtesting matters before trusting a strategy. I'm still experimenting, learning how the platform works, testing strategies and understanding automated trading. Not claiming to be an expert. Just documenting the journey. More experiments, more lessons. 👀 If you're curious, check out the platform: zoqo.trade/ @ZoqoTrade
🤝 Paid partnership
54
8
59
499
AI can summarise research quickly, but a fast summary can still mix verified facts with claims that aren’t fully supported. The useful part is being able to tell which points are backed by the sources and which ones still need checking. Have you caught an AI research summary that sounded solid but included something the original sources didn’t actually say?
10
4
24
562
Almusty Labs retweeted
An AI app should never hide an API failure. If the request hits a rate limit, auth error, or timeout, show it clearly instead of returning a half-finished or made-up answer. Silent failures are harder to debug than loud ones. What’s the most frustrating silent failure you’ve seen in an AI tool?
16
5
25
432