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)
QAFILA 2026: A Million Reasons to Celebrate with Binance! Binance is bringing its community together with a $1 million total reward pool through Qafila 2026. Running from October 20 to November 30, this campaign features different activities designed to give eligible users more ways to explore the Binance ecosystem. Here are some of the activities to discover: ➤ First Trade: Complete a qualifying first trade with at least $25 in trading volume and potentially receive up to 20 USDC. ➤ TradFi Perps: Explore eligible TradFi Perpetual Futures activities with a minimum $500 trading volume for up to 15 USDC. ➤ Trading Challenge: Participate in eligible Spot and Futures competitions and compete for a share of the 75,000 USDC reward pool. ➤ Simple Earn: Subscribe with at least 20 USDT and maintain the subscription for two days to qualify for 1.5 USDC. ➤ Referral Challenge: Invite friends and participate in qualifying referral activities for potential rewards. Why Qafila is worth exploring I like how the campaign brings different activities together instead of focusing on just one aspect of the Binance ecosystem. Whether it's trading, exploring Earn or connecting with other community members, there's something different to discover. Remember, activities and rewards depend on your region and eligibility. The $1 million is the total campaign reward pool, not a guaranteed individual reward. Explore the official campaign page: binance.com/en/ti/act/Qafila Check the available activities and full Terms & Conditions before participating. Which Qafila activity interests you the most? Disclaimer: General announcement, not financial advice. Digital assets involve market risks. Products and services may not be available in all regions. Review Binance's Terms of Use and Risk Warning. #Qafila #BinanceUpdate #Binance� #BinancePakistan
🤝 Paid partnership
174
73
1
296
4,327
An API is just a request and a response between your AI app and another service. You send a task, the service processes it and sends data back, then you inspect that data before using it. The basic flow is simple, until the request fails or the response isn’t what you expected. What’s an API issue that has broken an AI workflow for you?
7
13
270
Almusty Labs retweeted
A local AI can organise docs, code, and notes for a blockchain project pretty well. It still can’t prove anything. If a claim in the summary matters, you have to trace it back to the original source. Have you used an AI to summarise project docs and then had to correct something it pulled?
11
3
1
23
819
Almusty Labs retweeted
Quantising a model saves memory, but it doesn’t always keep the same quality. Some tasks degrade more than others. The only reliable way to know is to test the quantised version on the actual work you care about, not just assume it’s close enough. Have you noticed a clear quality drop after switching to a smaller quantised model?
8
2
16
713
Almusty Labs retweeted
This only further proves how terrible human drivers are and how we need to speedrun legalization of self driving cars worldwide
Surveillance video shows the crash that involved two Waymos and two cars at 20th St and Van Buren. @PhoenixPolice said 3 people were taken to the hospital: 2 had non-life threatening injuries and one was transported with life-threatening injuries. @abc15
54
22
4
565
53,668
Almusty Labs retweeted
I've been exploring new ways to sandbox code written by coding agents - here's a description of the external sandbox pattern, where agents run on your Mac but execute code in a WebAssembly or QEMU or Apple Container sandbox - with demos for Claude Code, Codex and Pi blog.primeradiant.com/extern…
213
62
32
779
40,634
Almusty Labs retweeted
Me, watching the first cut for the fifth time: The product looks good. The pacing is fine. But why doesn't it feel like a Black Friday campaign yet? A little creative flexibility helps. And if you're eligible, there's a desktop-only CapCut Pro offer with 90% OFF the first month on a monthly subscription. Eligible new users, users with expiring subscriptions, and former subscribers whose subscriptions ended more than 30 days ago can get the first month for $3.99 or less. Some new users may also qualify for a free trial. The right atmosphere can come from the footage itself, not just extra effects. Seedance 2.5 4K generates additional 4K AI video clips in CapCut PC through AI Video in the multi-track timeline or Video Studio. It's available in select regions outside the US, with the US rollout coming soon. Once those visuals are in place, music, transitions, and carefully timed text can turn a simple product showcase into a more recognisable seasonal campaign. Give your product showcase a stronger Black Friday feel: tinyurl.com/2v23rn22 #CapCutMarketing #CapCutPro90Off #CapCutPC #ChatwithCapCutAgent #CapCutDidThat #CapCutSeedance254k
28
23
29
507
Almusty Labs retweeted
Very cool remake for all the 🇵🇹 Portuguese that are now 25 to 30 years old who in ~2008 all got this laptop from the Portuguese government It was a branded version of Intel's Classmate PC with lots of open source software running on it like OpenOffice Kinda like an early Chromebook!
I just built an iconic Portuguese laptop simulator that runs in your browser! Let me explain: "Magalhães" was this "Portuguese classic" - a low-spec Linux-based laptop that the Portuguese government basically "gave" to a ton of students around 2008-2010 (its price ranged from 20€ to 180€ depending on your household income) For many young Portuguese in their ~25-ish, playing SuperTux (an open-source game btw) is deeply nostalgic People would literally install games on this 'educational laptop' like Metin 2, CS 1.6, and more. It sure powered a generation of gamers 🤣 Anyway, you can now play SuperTux, paint, create documents, pptx slides, play music, and I might just add browsing later on (imagine claude/chatgpt ahah) Try it now! magalhaes-simulator.vercel.a… (needless to say, heavily inspired by @levelsio's pieter.com ) x.com/goncalo_canhoto/status…
8
5
1
134
26,053
Almusty Labs retweeted
A chatbot can undo a mistake. A robot that spills water can't. So what is physical AI? In 3 minutes: the sense-decide-act loop, the 5-layer robot software stack, why modules now learn from data, and the 3 routes to controlling robots with large models (VLA, world models, WAM).
5
2
32
812
Quantising a model saves memory, but it doesn’t always keep the same quality. Some tasks degrade more than others. The only reliable way to know is to test the quantised version on the actual work you care about, not just assume it’s close enough. Have you noticed a clear quality drop after switching to a smaller quantised model?
8
2
16
713
Almusty Labs retweeted
Notes from the October meeting drop this week. Curious what the 2 who wanted a hike had to say
3
6
101
Almusty Labs retweeted
A local model is useful for organising your own notes and research. It’s not useful for live market data unless you still pull that from somewhere else. Most people mix the two and then wonder why the answers are outdated. Are you using a local model mainly for private notes, or are you also trying to get it to handle live data?
12
1
19
852
A local AI can organise docs, code, and notes for a blockchain project pretty well. It still can’t prove anything. If a claim in the summary matters, you have to trace it back to the original source. Have you used an AI to summarise project docs and then had to correct something it pulled?
11
3
1
23
819
A local model is useful for organising your own notes and research. It’s not useful for live market data unless you still pull that from somewhere else. Most people mix the two and then wonder why the answers are outdated. Are you using a local model mainly for private notes, or are you also trying to get it to handle live data?
12
1
19
852
One campaign. Different opportunities. A million reasons to celebrate. That's what makes Binance Qafila 2026 interesting to me. With a $1 million total reward pool, Qafila brings together different activities for the Binance community, running until November 30, 2026. Here's what caught my attention: 📌 Trading: Start with your first eligible trade and qualify for up to 20 USDC. 📌 TradFi Perps: Explore eligible TradFi Perpetual Futures trading for a chance to earn up to 15 USDC. 📌 Trading Competitions: Put your Spot and Futures trading skills to the test and compete for a share of 75,000 USDC. 📌 Simple Earn: Subscribe with at least 20 USDT, maintain it for two days and qualify for 1.5 USDC. 📌 Referrals: Invite friends to join the journey and explore eligible referral rewards, including leaderboard opportunities. What I appreciate most is that Qafila brings different interests together. Some people enjoy trading challenges, others prefer exploring Earn, while some are more interested in building connections within the community. There's no need to rush or participate in everything. Understand the requirements and check which activities are available in your region before getting involved. The $1M represents the total campaign reward pool, not a guaranteed individual reward. Explore Binance Qafila 2026: binance.com/en/ti/act/Qafila If you could choose one activity, which would it be: Trading, Simple Earn or Referrals? Disclaimer: General announcement, not financial advice. Digital assets involve market risks and price volatility. Products and services may not be available in all regions. Eligibility and rewards are subject to the official campaign Terms & Conditions. Always review Binance's Terms of Use and Risk Warning. #Binance #Qafila2026 #BinanceSquare #CryptoCommunity
7
13
672
Almusty Labs retweeted
Running the model locally doesn’t mean your data stays local. A lot of setups still send prompts or logs to external APIs, telemetry services, or store them in places you didn’t check. The model location is only one part of the flow. Have you found an app that was running a local model but still leaking data somewhere else?
10
3
19
1,278
Almusty Labs retweeted
You will never be profitable without God on your side
23
39
4
993
12,540
Almusty Labs retweeted
I'm not cancelling ChatGPT just to have Claude code me a fucking cat selfie.
407
180
81
8,853
3,219,003
Almusty Labs retweeted
A review gets clearer when taste and task stop sharing one label. I prefer blue is a different kind of comment from I cannot find the booking action. Both can be discussed, but they need different responses. This set ends with a proposed usability question, not a claim that a test has already been run. Name the goal, identify the concern and choose how it will be checked. That makes the next version easier to explain and keeps a review from becoming a contest of confident preferences. Useful feedback gives the work a direction it can actually follow. Explore my services: mrcreatech.site
5
3
22
290
Almusty Labs retweeted
🌎 My Weekly Founder's Coworking Day is going global! I'll be in 🇸🇬 Singapore and 🇹🇭 Bangkok soon, and I'd love to host my weekly coworking day for founders there! If you're interested to join let me know, and if you know a good venue also 🤝 😊
I've been actively fighting almost guaranteed age-related loneliness for years, studies show after age 30 you start seeing friends less and when you get older especially men don't have any social life left! It's preventable, but it takes effort and initiative We've now organized 70 weekly coworking days where we meet with friends and work together, do show and tell, and try help each other make things (also hardware stuff more recently) We've also started organizing movie nights, art nights, barbecues, we watched the World Cup final on a projector screen in the garden with friends We did a collage night the other day where we bought cheap gossip magazines, colored paper, and Pritt glue, and gave everyone scissors to make their own art, I think nobody spent time on their phone for at least 4 hours, making things puts you in a flow state and it's fun to do it in a group because you talk about different stuff while making things than you would for ex having a dinner Movie nights are nice too but less popular I think cause everyone prefers to watch movies alone and also you can't really talk during a movie Coworkings range from having sometimes 20 people to sometimes just 2 people, people have their own lives, but me and @rameerez still try to organize them weekly One thing I realized is that we are usually the ones organizing stuff, but that's not a bad thing, there's even studies about this, there's usually a few people in a group who organize most of the events If nobody in your group organizes stuff, it doesn't mean nobody wants to do stuff, but you have to take the initiative to bring people together which isn't easy in a time where phones and apps and the internet are simply easier and cheaper dopamine than driving to a friends house for a party You can start with a coworking day, everyone brings their laptop, and you provide WiFi and coffee, then later you can try more types of events Having an active social life is just as important or in some studies even more important for health than physical fitness After a nice party I feel tired but satisfied, I had fun and didn't spend time on my phone or computer but instead talked and laughed with friends etc. We don't drink alcohol with very few exceptions and nor do our friends, so it's mostly just having really nice food and sparkling water or kombucha etc, and we don't really need alcohol, after a few hours it gets fun enough anyway! So I recommend everyone, take initiative, and organize something and get people together, it'll be fun! 😊
198
13
8
477
47,083
Running the model locally doesn’t mean your data stays local. A lot of setups still send prompts or logs to external APIs, telemetry services, or store them in places you didn’t check. The model location is only one part of the flow. Have you found an app that was running a local model but still leaking data somewhere else?
10
3
19
1,278