Joined December 2023
- Tweets 397
- Following 169
- Followers 52
- Likes 380
329 Photos and videos
Tegaki: a handwriting animation library for the web. Works with any font/text. Includes a stroke data generator. Supports React, web components, vanilla JS and Remotion. Control via code or CSS scroll timelines. Handles multiline text
#WebAnimation #Handwriting #FE
1
67
Someone built WeClone that reads all your chat messages and creates an AI clone of you. Export chat history from WeChat Telegram or any app. It fine-tunes an LLM on your real conversations capturing your slang humor tone and personality.
#WeClone #AI #DigitalTwin #OpenSource
43
Hermes Agent v0.9.0 The Everywhere Release is out. Local web dashboard, Android Termux support, iMessage, background monitoring, backup/import, and more. Update with hermes update. Full notes: github.com/NousResearch/herm…
#HermesAgent #AI #OpenSource
73
OpenRouter welcomes a new stealth model: Elephant Alpha. A 100B parameter instant model matching SOTA performance of similar scale while extremely token efficient. Strong at code completion debugging document processing and lightweight agents
#ElephantAlpha #OpenRouter #LLM #AI
320
Mistral just added built-in transcription for meetings interviews and more with speaker diarization and segmentation. Convert audio to accurate structured text effortlessly. Try it here: v2.auth.mistral.ai/login?flo…
#MistralAI #Transcription #SpeechToText #Diarization
8
Google open sourced Magika: their secret AI weapon for real file detection. Ran internally on hundreds of billions of files weekly across Gmail Drive Safe Browsing. AI trained on 100M files, 200 types, 99% accuracy, 5ms per file
pip install magika
#Magika #Google #AI #OpenSource
1
19
Lovable just launched Lovable Payments. Describe what you want to sell. Test securely. Set up payments in one minute, ask about your payments data, VAT/sales tax and currency conversion for 200 countries. Integrates with Stripe, Paddle or Shopify.
#LovablePayments #AI #SaaS
2
1
62
Cursor 3 just got several quality-of-life improvements. Split agents for multi-tasking like in your terminal, more reliable voice input, jump from diff straight to file line, choose branch for cloud agents, workspace search with filters
13
Google Research introduces ConvApparel: new human-AI conversation dataset in fashion retail plus an evaluation framework to quantify the realism gap in LLM user simulators. Aims to build more robust conversational agents. Read the blog for details.
#ConvApparel #ConversationalAI
22
Building a mobile game from scratch with Autodesk Flow Studio. All assets generated with Wonder 3D from blocks to bulldozers. A faster way to turn an idea into a playable world. What world should we build next?
#Wonder3D #Autodesk #GameDev #AI3D #MobileGames
3
33
Gemini can now transform questions and complex concepts into customizable interactive visualizations directly in chat. Adjust variables rotate 3D models and explore data for immersive learning. Try it on gemini web with Pro model using show me or help me visualize prompts
40
#Perplexity Computer now connects with #Plaid to link bank accounts credit cards and loans. Track spending in detail build custom budget and visualize net worth with your investment portfolio. All users can link accounts and ask about finances. Computer tasks for Pro and Max only
30
Claude is bringing the advisor strategy to the platform. Pair Opus as advisor with Sonnet or Haiku as executor to get near Opus-level intelligence in agents at a fraction of the cost. When the agent hits a hard decision it consults Opus gets a plan and continues all in 1 API req
23
Anthropic launches Advisor Strategy: Use Opus as advisor Sonnet/Haiku as executor for near Opus-level performance at lower cost. 2.7% on SWE-bench while costing 11.9% less. Now in beta.
#Claude #Anthropic #AI
16
Mistral just added realistic speech generation from audio references. Record or upload your voice → control tone, rhythm, intonation & emotion. Impressive emotional expression! 🎙️
#MistralAI #VoiceAI
2
Meta launches Muse Spark – the first model from our Superintelligence Lab. Smarter reasoning, multi-agent thinking, efficient test-time scaling. On a predictable path toward personal superintelligence. Excited for what's next!
#MuseSpark
14
Genspark AI now natively embedded in Microsoft Office as plugins! Create slides from prompts in PowerPoint, analyze data & generate charts in Excel, write polished docs in Word — all inside the apps you already use. No switching tools.
#GensparkOffice #GenSpark #Office
82
Paperdraw: Drag-and-drop system design simulator that shows real-time behavior under traffic, failures, latency & scaling. Great for learning, validating architectures, or teaching. 100% free at paperdraw.dev.
#Paperdraw #SystemDesign
18