The AI world moves fast. We bring you the moments that matter. Tools ✦ Tips & Tricks ✦ Insights ✉️ Biz [email protected]
California Usa 🇺🇸
Joined September 2026
- Tweets373
- Following222
- Followers108K
- Likes544
Maya retweeted
AI is no longer just answering prompts. It's starting to handle entire workflows.
Ling-3.2-flash can turn documents into code, images into 3D scenes, financial data into presentations, and ideas into interactive websites.
One multimodal model covering so many different tasks is genuinely impressive.
Introducing Ling-3.2-flash, Ant Ling's native multimodal model: 560B total / 25B active parameters.
Stronger complex visual understanding, long-horizon multimodal agents, 3D understanding & interaction, and medical intelligence.
Outscores Opus 5 (max) in release evaluations:
- OmniDocBench1.5: 92.21 vs 90.56
- ZEROBench(Sub): 43.71 vs 39.52
- Vision2Web: 85.86 vs 80.20
Maya retweeted
Re-recorded one line. Now my subtitles disagree with me, and the new take is twice as loud. Tiny edit. Huge personality.
Tidy up your next ad in CapCut PC.
➪ tinyurl.com/2v23rn22
#capcutpc #capcutblackfriday
🤝 Paid partnership · 🤖 Made with AI
Maya retweeted
solar panels are the demo but think about warehouses, factories, defense ops
every one of those robots hits stuff nobody wrote a test for. @Zynoptes builds those situations as simulated worlds and fixes what breaks in the code.
Break your robot before the world does.
Introducing @Zynoptes (ALL18), an Autonomous Simulation Engineer for Robotics that finds and fixes issues before robots reach the customer's site.
You bring your robot’s task and code. Zynoptes builds the simulations, runs hundreds of scenarios, and validates code fixes so your team can confidently deploy.
Congratulations on the launch, @theOwenBales and @KernMaxim87213!
Maya retweeted
AI assistants are becoming more interesting when they can help with real-world tasks.
Well brings that experience directly to WhatsApp, helping users manage emails, receipts, and everyday business tasks from one place.
🤝 Paid partnership
Aujourd'hui, on lance Well en France 🇫🇷
Un assistant IA qui prépare l'administratif de ton entreprise, directement sur WhatsApp.
Être entrepreneur en France, c'est aussi avoir un deuxième métier à plein temps : gérer l'administratif.
Et ça ressemble souvent à ça :
📥 245 emails à gérer entre clients, fournisseurs, URSSAF, impôts, banque et comptable
🤝 56 engagements pris en réunion : relances, devis, rappels, rendez-vous
🔐 37 outils entre lesquels jongler : Qonto, Pennylane, Stripe, HubSpot, Google Agenda, impots.gouv, URSSAF…
📚 10 400 pages de lois et de règlements que nul n'est censé ignorer :
Code du travail, Code général des impôts, Livre des procédures fiscales, CIBS, Code de commerce, Code de la sécurité sociale
Et tout ça avant même de commencer à travailler sur ton vrai métier.
C'est exactement pour ça qu'on a créé Well.
Pas de nouveau logiciel à apprendre.
Pas de tableau de bord à consulter.
Juste un message à envoyer.
☀️ Avec Well, tu prends ton café, tu envoies un message et tu reprends le contrôle de ta journée.
Toi, tu fais ton métier.
Well prépare le reste.
Essaie Well aujourd'hui 👇
wa.me/13025202443?text=Hello…
Maya retweeted
"I’m giving my product a UGC ad with an AI presenter doing the demonstration. CapCut PC is where I’d build it. Friendly delivery, clear product shots that’s the idea.
I start in Design Studio, then use Video Studio to turn the presenter and product setup into clips. The preview adds a humming noise and a wall stain I absolutely did not need. Now I’m inspecting the wall instead of the demo.
Through Edit Pilot, I repair the patch with AI Edit and clean up the audio with Reduce Noise and Enhance Voice. I check the changed area during playback.
Give your product demo a cleaner finish in CapCut PC.
➪ tinyurl.com/2v23rn22
#capcutpc #capcutmarketing #capcutblackfriday #CapCutDidThat
🤝 Paid partnership · 🤖 Made with AI
Maya retweeted
AI collaboration is an interesting frontier because it changes the unit of productivity—from one capable assistant to a coordinated team of agents and people.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Maya retweeted
I've always wondered why making a game with AI still requires jumping between so many different tools.
Wanaka seems to be taking a different approach by bringing the entire workflow into one place. Curious to see how @wanakaofficial pulls it all together.
Introducing Wanaka Studio 1.0, an AI-native game engine.
Build real 3D multiplayer games. No coding needed.
Instead of making AI fight with Unity, we built an engine it can actually use.
1️⃣→ All-in-one workspace
3D models, rigging, gameplay logic, and sound.
2️⃣→ Speed & cost
5x faster than Opus 5.5 + Unity, at 82% lower agent cost (benchmark in thread).
3️⃣→ Performance
AAA-style visuals in 4K, 60+ FPS, real-time physics and lighting.
4️⃣→ Direct control
Edit anything right in the scene. The agent picks up from your changes.
5️⃣→ One-click publish
Every game is just a link. Cross-platform, nothing to install.
Try Wanaka → studio.wanaka.app/
Maya retweeted
Your OCR says it’s done—but did it get everything right? 🤔 In high-stakes workflows, even a small error can be costly.
🚀 Meet the MiniCPM-OCR suite: an open-source solution to parse fast and guard the results.
Two new models:
🔹 MiniCPM-OCR 1.0 — SOTA OCR performance and speed at 1B scale. Compact, accurate, and fast.
🔹 MiniCPM-OCR Jev — A 2B decision model that helps assess parsing accuracy, with performance approaching that of Qwen3.5-397B-A17B on JevOCRBench.
⚡ What’s new?
Small model, SOTA results: MiniCPM-OCR 1.0 scores 93.68 on Real5-OmniDocBench and 54.7 on OCRBench v2.
A sharp goalkeeper 🛡️: MiniCPM-OCR Jev is model-agnostic, checking text, formulas, and tables from different OCR parsers. It achieves 93.6 on JevOCRBench, outperforming JEV-35B-A3B (90.1) and approaching Qwen3.5-397B-A17B (93.8).
Up to 128 pages per minute 🤯: With DSpark, MiniCPM-OCR runs at 1.49× the speed of PaddleOCR-VL-1.6 and 3.62× that of MinerU2.5 Pro. The SDK’s concurrency optimizations further boost throughput to as much as 128 pages per minute.
Deploy where you work: Day-0 support for Arm, Houmo AI, Intel, and NVIDIA, plus 10 more chip vendors through FlagOS.
Ready for your stack: vLLM, SGLang, llama.cpp, and Ollama, with quantized models, fine-tuning support, and Agent Skills.
🔗 GitHub: github.com/OpenBMB/MiniCPM-O…
🤗 Hugging Face:
MiniCPM-OCR 1.0：huggingface.co/openbmb/MiniC…
MiniCPM-OCR Jev：huggingface.co/openbmb/MiniC…
👉 Try the Demo：platform.modelbest.ai/consol…
Use the API：platform.modelbest.ai/consol…
Maya retweeted
Agents f*ck up.
When your SDR agent does, it shouldn't take the domain your invoices go out from with it.
OpenMail gives every agent its own inbox, its own sending limits and its own reputation.
openmail.sh