Ismail Khan (@learnbyIsmail) header
Ismail Khan (@learnbyIsmail) profile picture

🤖 AI Automation & Tools 🎥 Sharing Full Courses & Tutorials 🚀 Helping You Learn AI & Automate Smarter [email protected]

Georgia, USA
Joined November 2020
Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)Photo by Ismail Khan (@learnbyIsmail)
Ismail Khan retweeted
My moodboard made perfect sense until I put it all in one prompt. Turns out I’d saved four different aesthetics under the same name. The CapCut Web Ultra event is live. From Nov 13–19, buying or upgrading to Ultra gets you 7 days of unlimited image generation on Seedream (5.5 / 5.0 Pro / 5.0 / 4.5 / 4.3) and GPT Image 2.5 at 1K or 2K. Available in the UK, France, Germany, Italy, Spain, Mexico, Brazil, Canada, Japan, South Korea, and Australia. capcut.com/my-edit?utm_mediu… #CapCutWeb
🤝 Paid partnership · 🤖 Made with AI
22
72
179
75,810
Replying to @HyperHaulio
@hyperhaulio hyperhaul.io/waitlist?ref=HA… It's looks an alpha opportunity don't miss this. Top 300 will get the rewards let's try to crack the leaderboard
14
24
94
Ismail Khan retweeted
A moving camera shouldn’t mean a moving world Obvious to a person, surprisingly hard for a model Here’s what I got from PAN 2 on early access: a flight around a peak as first light hits it, and the mountain keeps its shape
🤝 Paid partnership
Introducing PAN 2: an open foundation world model for physical intelligence. A language model predicts the next word. PAN 2 predicts the next world: what happens after each action, grounded in time, physics and space. Towards AI that simulates “what if?” before it acts. More details: ifm.ai/pan/
29
20
152
91,366
Ismail Khan retweeted
MiniMax H3 Image is now available for early access on Hailuo AI. What caught my attention is the support for up to 16 reference images. That could be really useful when trying to maintain a consistent look while exploring different creative directions. From character designs and concept art to 3D environments, there's a lot to experiment with. Curious to see what creators build with this one. @MiniMax_AI #H3Image #AIImageGeneration
🤖 Made with AI
21
23
147
29,635
Ismail Khan retweeted
🚨 You can win every deal just by using AI right on your sales calls. Backdrop helps you during the call. It listens quietly and shows you what to ask or say next, right when you need it. Stop losing deals. Start winning them….
🤝 Paid partnership
Selling is hard. Live calls are brutal. Forget one question and the deal might be gone. Your buyers are impatient. Someone pitches them a version of your product every single day. Until now, AI couldn't help you in the moment. Too slow. Too distracting. Not anymore. Introducing Backdrop: the first live sales assistant that doesn't suck. Rookie or veteran, it doesn't matter. Backdrop has already helped dozens of teams close 6-figure deals.
4
29
174
53,813
"Everyone can be a pretty good marketer if you ask Claude or if you ask ChatGPT." @bokelley of Apostra on why breaking through still takes a creative spark no model hands you. Six takeaways on the Craft blog: craftventures.com/articles/m… Follow @unpromptedcraft to get every session early and first word on the next one while spots last.
10
106
976
Ismail Khan retweeted
Every voice AI company will try to come up with a Duplex model in 2027. It's basically a single model that: - Listens continuously while speaking - Decides when to speak or pause - Handles interruptions and corrections smoothly - Uses backchanneling like “mm-hm” or “right” - Streams speech in real time - Adapts its reply mid-sentence - Manages turn-taking naturally Every other agent is just a hoax made up with 4 different models
🤝 Paid partnership
Introducing @FishAudio Live 3: a full-duplex voice model that listens, says "mm-hmm," and knows when to shut up. Enterprise? Book a demo for your free custom agent: s.fish.audio/vxhgs6 Reply with your use case - we'll build it free for the first 1,000 people.
15
23
134
40,500
Ismail Khan retweeted
Tamil, Telugu and Malayalam each make about 30% fewer word errors on Scribe v3 than on v2. Going to test Hindi first and see how accurately it handles everyday conversations.
🤝 Paid partnership
Introducing Scribe v3 and Scribe v3 Live: our most accurate Speech to Text models yet. Built for real-world audio, Scribe excels in situations with extreme background noise, knows pauses from the end of sentences, and keeps track of who said what in a busy meeting.
17
29
121
36,845
Ismail Khan retweeted
state drift between our agents is our worst bug, and it mostly comes from a human copying context between them by hand. in Teamily every agent reads off the same group context and people post text, screenshots and files right into it, which is most of what we hand sync now.
🤝 Paid partnership
Messaging Connected People. AI-Native Messaging Will Connect the Agentic Internet. The first generation of instant messaging connected people. It changed how billions of us communicate, socialize, and work together. But the next generation of the Internet won't be built for humans alone. AI agents are becoming our teammates, creators, and service providers. To truly work alongside humans, they must communicate and collaborate as naturally as people do, while going far beyond human capabilities with richer context, living memory, continuously self-improving intelligence, and seamless coordination across people, agents, and services. Yet today's messengers were designed for human-to-human communication, not human–agent collaboration. They lack the native capabilities to support intelligent, context-aware human–agent teams. It's time to reinvent messaging for the Agentic Internet. That's why we're reinventing instant messaging at Teamily AI. 👥 People and AI agents working together in the same conversations. 🧠 Unified context and living memory across people, agents, and services. ⚡ Conversations that turn into real actions, creations, and automations, as well as agentic websites and apps as a service. 🌐 An open gateway where every agent and service is as accessible as messaging a friend. Traditional IM connected people. AI-native IM connects people, agents, and services. This is how we move from the Internet of Apps to the Agentic Internet. And we're just getting started. teamily.ai #TeamilyAI #AgenticInternet #AINativeIM #TeamAgent #AgenticIM
8
35
122
49,164
Ismail Khan retweeted
EinsiaAI's WorkflowArena grades the merge at the end too. that's where I'd expect a lot of runs to fall apart, when the lead agent has to pull everyone's work back together
Agent harnesses now let a model spawn its own workers and write the workflow that coordinates them. But can it actually orchestrate them well? Introducing Einsia AI's WorkflowArena, the first benchmark for autonomous multi-agent workflow orchestration on real, long-horizon work. - 145 tasks across 8 professional domains, from SWE and Lean proofs to law and 3D - 25 configurations of 9 frontier models in Claude Code & Codex - Every delegation recorded: we score the deliverable and see the workflow behind it 🌐 Project: lab.einsia.ai/workflowarena 💻 GitHub: github.com/Einsia/WorkflowAr… 📄 Paper: arxiv.org/pdf/2611.12856 ⬇️ Read the following threads to go through the key findings
7
12
93
91,932
Ismail Khan retweeted
You can’t train human-level manipulation without human-level viewpoint data. Memories.ai is systematically solving this data shortage.
🤝 Paid partnership
Today we launch #MemoriesUpload. Robots need internet-scale first-person video of how people get things done. Collecting this at high quality takes hardware, model and infra working together. That's what we've spent over a year building. And already, 3 of the top 5 frontier robotics labs are training models with the data.
7
11
68
42,371
Ismail Khan retweeted
Ever stare at a finished promo video and think... why doesn't this feel like Black Friday yet? You've got the flyer flying into the frame, but the overall design still feels like an ordinary announcement. Sometimes you don't need to redesign everything. CapCut PC has Black Friday stickers, effects, and text templates that help give existing footage a more seasonal look and make promotional offers stand out. A little festive treatment goes a long way. Here's where I'd give my promo a Black Friday makeover: tinyurl.com/2v23rn22 #capcutpc #capcutmarketing #capcutblackfriday #CapCutDidThat #capcuteditpilot
🤝 Paid partnership
22
18
140
75,438
Ismail Khan retweeted
Imagine your AI meeting assistant transcribes everything perfectly. But assigns your boss's instructions to the wrong person. THAT'S WHERE THINGS GOES WRONG! AirCaps just launched A5S Async v1, built to get both the words AND the speakers right. It's #1 of 15 on VoiceArena. Developers can try it free for 5 hours.
🤝 Paid partnership
Introducing A5S Async V1: the world's most accurate speech recognition model for speaker-attributed transcription. #1 on VoiceArena, ahead of ElevenLabs, AssemblyAI, Deepgram, NVIDIA, Meta & more. The next-best model makes 41% more errors across 5 public benchmarks. And yes, every lab says they're #1. So we open-sourced our benchmark. Try it now👇
22
13
82
21,496
Ismail Khan retweeted
speaker diarization only gets tested when people start interrupting each other.. so i made it harder. 3 astronauts in identical suits with mirrored visors, so you can't see who's talking. the voice is the only clue i ran the 45s scene through MOSS-Transcribe-Diarize Pro by @MOSS_Global ➤ 28 lines, split into S01, S02, S03 by voice alone ➤ it caught the interruptions and the moments two people talk over each other ➤ start and end timestamps on every line ➤ works on the platform or through the API, with keyterms for custom names every line landed on the right speaker.. full demo in the video. try it here: platform.moss.cn/?utm_0=2mIB…
25
20
197
64,269
Ismail Khan retweeted
Bookmarking this so I stop skipping the planning step.
🤝 Paid partnership
You’ve spent the past hour generating a video, and it still looks like AI. You’re probably making one of these 6 mistakes. These fixes will get you more realistic results in fewer generations. Bookmark this for your next project.
7
21
87
52,674
Ismail Khan retweeted
One creative concept can go further than a single campaign. Start with someone holding a stack of papers, let a sudden gust scatter them, then have one flyer fly toward the camera to reveal the product announcement. It's a fun format to reuse, but changing the campaign details every time can get surprisingly repetitive. This is a workflow worth exploring in CapCut PC: tinyurl.com/2v23rn22 #capcutpc #capcutblackfriday
17
18
111
85,207
Ismail Khan retweeted
Last week I added Viktor to my org chart. This week he got the official team announcement. One thing that has changed since then is how I manage brand collaborations. Between incoming campaign briefs, draft reviews, content deadlines, and follow-ups, it's easy for small but important details to get lost. Recently, Viktor pulled together my campaign updates, tracked pending draft reviews, and flagged the brand follow-ups that needed my reply. This saves me around 2 hours every week that I would otherwise spend manually checking multiple conversations and tracking updates. He pulls the campaign updates together and flags what needs my reply. He also helps me keep track of pending drafts and follow-ups, so nothing important slips through the cracks. It's a much smoother way to manage the day-to-day work behind my creator business. I'm still in control. He proposes, I approve. Still a one-person team. Just not doing it all alone anymore. Hire @viktor_com for your team. Start free. #AIemployee #AItools
🤝 Paid partnership
32
24
188
27,517
Ismail Khan retweeted
Building Complex Procedural 3D Scenes with Three.js | Powered by Hy4 Exploring the creative and technical capabilities of Hy4 through WorkBuddy by transforming a simple, one-line prompt into a sophisticated procedural 3D scene using Three.js. The project demonstrates how AI can translate natural-language instructions into complex 3D experiences, combining procedural generation, programming, and visual design. Key Highlights → Procedural 3D scene generation → Complex geometry and environmental composition → Rich visual details and spatial depth → From a single prompt to a complete 3D experience Why it matters: Hy4 demonstrates promising potential in bridging the gap between creative vision and technical implementation, making complex 3D development more accessible through natural-language interaction. One prompt. Complex 3D creation. Powered by Hy4. You can explore WorkBuddy here: workbuddy.ai @TencentHunyuan @TencentAI_News @WorkBuddy_AI
24
16
121
58,716
Ismail Khan retweeted
most transcription tools give you a wall of text, and you still have to figure out who said what.. MOSS-Transcribe-Diarize Pro by @MOSS_Global does that part for you i tested it on a 60s clip with 3 speakers talking over a loud battle in the background ➤ it labeled every voice as S01, S02, S03 by itself ➤ each line came with a start and end timestamp ➤ keyterms let me add custom names so they're spelled right ➤ it runs on the platform or through the API, so you can build it into your own app 29 lines, 3 speakers, no manual cleanup.. full demo in the video. try it here: platform.moss.cn/?utm_0=2mIB…
31
23
210
89,889
Ismail Khan retweeted
Ever wonder why some videos get millions of views and yours don't? this guide answers that.
🤝 Paid partnership
This entire video was made with Seedance 2.5 and SFX v3. Learn how to make your own and what gets it picked up by the algorithm. Save the full guide below.
7
24
189
91,968