Abdul (@dev_ludba) header
Abdul (@dev_ludba) profile picture

Building OS stuff....

Joined November 2024
Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)Photo by Abdul (@dev_ludba)
Winner announcement in 24 hours! We'll randomly draw 100 eligible winners with help from @grok. Make sure you follow us, like & repost to be in the running. Good luck! 🍀
My iPhone Duo video went viral, so I'm giving away 100 iPhones on X 😛
13,827
17,799
966
25,482
1,447,499
My iPhone Duo video went viral, so I'm giving away 100 iPhones on X 😛
94,245
96,577
6,673
143,820
5,496,446
Shits about to get really Islamophobic.
This Post is from a suspended account. Learn more
13
1
29
1,331
Abdul retweeted
Grok Bot can now do all your email marketing Introducing our official integration with Grok Bot 🎉 Design, segment, and send high converting email campaigns natively.
93
58
20
297
1,032,797
We started telling our own users to not use our App Builder. Why? Because as long as they use their own Claude Code (or Cursor / Codex / etc), they're better off that way. It is almost ridiculous seeing the number of App Builders in the market. We built ours with a focus on complex applications, but our users kept telling us they prefer their own tools once they build something complex. So we ended up letting you bring your own Claude Code and build without buying AI credits - going forward you'll see a "Continue in Claude Code" button in Modelence. Btw, a lot of other platforms are now doing something similar, and we deliberately do this for coding agents and not your plain Claude / ChatGPT. We tried that first and weren't happy with the result - a plain chat with MCP is not enough, you need proper coding tools with files and diffs to build complex applications. We are focusing more than ever on production infrastructure. After running our App Builder for over 7 months, it's clear that deploying, scaling and running production infrastructure is the biggest challenge everyone needs to be solved for them once they are building a serious app.
1
2
5
173
Abdul retweeted
💍💍Personal announcement: Migma is now engaged to Grok Bot. The wedding is coming soon
15
8
3
60
1,680
brand guidelines 💅
3
1
1
62
3,100
Ran out of space in the round, had an investor ask me yesterday > "so how do i find people like you earlier" < "how are you filtering" > "you know, ex stanford, harvard, mit, big tech" < "that's the problem" All the best founders i know have a massive chip on their shoulder This guy is quite literally filtering them out.
50
8
406
25,628
Abdul retweeted
damnn
1
6
261
Abdul retweeted
guess who powers payments for this beautiful product ;)
Today we're launching Endstack, the most capable agent workspace ever built One superagent, a collaborative desktop, and channels for code, automations, and ops. Out on desktop, web, and mobile We'll send you $100 if you can't automate your work with Endstack Rules below
1
5
205
Abdul retweeted
My SaaS didn't fail. My passport did. Solo founder. Lebanon. Every processor said no. @dodopayments said yes. For 5 months it finally worked. This morning they closed my account. Not fraud. Not chargebacks. Just the country on my ID.
57
3
1
282
39,816
Impressive gains across agentic benchmarks. 👀 Once the V4 Pro 0813 weights are released, we’ll work to bring it to OpenRouter as soon as possible. Meanwhile, Novita is already the #1 provider for DeepSeek V4 Flash 0731 by token share on OpenRouter, with a 25.3% token share and 95% cache hit rate. Try it through Novita. 🚀
DeepSeek V4 Pro 0813 is live on OpenRouter. @deepseek_ai reports large agent gains over V4 Pro Preview: DeepSWE 62.7 (+49.9), CyberGym 83.3 (+30.6), NL2Repo 61.5 (+23.0), and Terminal Bench 2.1 87.9 (+15.8) More providers coming online soon Use it now: openrouter.ai/deepseek/deeps…
4
1
1
32
3,734
Screenshotted a layout bug. Attached it. Typed one sentence about what was wrong. Fixed in 55 seconds. The first build is what gets demoed. The tenth small fix is what decides if you ship.
1
2
99
Abdul retweeted
✨👯 NO more AI slop in your social media Zeely generates winning Instagram carousels from your product link in minutes. For the next 69 hours RT, follow, comment ZEELY👇 to start with our viral carousel ideas and $100 promo code for new users.
45
37
2
65
50,595
Computer use with nothing installed. No SDK. No Playwright. No Docker. No browser driver. No VM to babysit. One POST. One field. One outcome. We provision the desktop, do the work, verify it, and destroy the machine. { "task": "Open the billing portal, download the newest invoice, and verify the PDF exists.", "max_steps": 150, "deadline_seconds": 1800, "webhook_url": " " }
2
3
8
529
AIXI Labs is accepting applications for its Quarterly Research Fellowship. This remote or London-based role supports AI safety research grounded in algorithmic information theory, offering a monthly stipend of £2,000–£3,500 plus a compute budget. Selected fellows will: - Lead self-directed projects on AIXI-style models or LLM-based agent tests. - Receive mentorship from Cole Wyeth, Marcus Hutter and Aram Ebtekar. - Focus on theory or empirical red-teaming. - Publish papers or blogs for the AI safety and ML communities. - Join a global network of researchers in UAI/AIXI and provable safety. To apply: - Submit via the AIXI Labs portal. - Demonstrate PhD-level research background (or Master's with substantial experience). - Show expertise in probability, statistics, information theory and reinforcement learning. - Highlight communication skills and motivation to reduce AI existential risk. - Optional: Publications at NeurIPS/ICML/ICLR/ALT/COLT, background in algorithmic information theory or LLM/RL agent experience. Applications are reviewed on a rolling basis... previous applicants remain eligible. Remote or London-based. If you work on provable AI safety or algorithmic information theory, apply here: aixi.uk/team/opportunities/a…
1
7
71
5,001
Abdul retweeted
⚠️ Bug bounty scam alert 🧵 Don't hunt on @huntr_ai (@PaloAltoNtwks). I found 14 critical/high vulns in Kubeflow. Huntr scoped the engagement to kubeflow/kubeflow a repo that contains ONLY a README. All my reports were closed NA and I got banned. 10k+ in bounties not paid!
16
19
3
243
19,694
WHAT THE FUCK This is real
1,074
22,932
707
131,534
2,875,326
Abdul retweeted
Kimi 2.7 Code now available in Go text · image · optimized for coding similar pricing as 2.6
105
143
36
3,934
173,423