@javaeeeee1i
iAccount based inCanada
About this account
- Account based in
- Canada
- Connected via
- Canada Android App
Account-level information from X, not a live location or the device used for a specific post.
Machine Learning Engineer. I learn by teaching. Opinions are my own. Building https://nitter.cf/t.co/LDTxX1BwWC
Toronto, Ontario
Joined October 2014
- Tweets19.9K
- Following1.7K
- Followers1.7K
- Likes12.6K
Building Scalable Apps with Firebase: A Comprehensive Developer's Guide to Tools, Technologies, and Real-World Applications @Firebase #firebase
reddit.com/r/AgentContext_de…
Dmitry Noranovich retweeted
Opus 5.5 is 20% cheaper per input and output token than Opus 5, and 60% cheaper on cache reads. So what does that actually do to the cost of a task in Claude Code?
We ran the numbers, and built a calculator so you can run yours from /usage:
claude.dev/blog/what-a-task-…
Dmitry Noranovich retweeted
It’s now easier to build plugins for Claude.
We built a new portal to submit your plugin, track review, and see usage.
Plugins package MCP and skills, and are becoming the way to build for Claude. MCP usage across Claude products is up 110x this year!
claude.com/blog/build-plugin…
Dmitry Noranovich retweeted
Reasoning from scratch, round number 5! This time, talking about log-probability scoring (also a great fundamental concept for loss functions like cross-entropy in pre-training and distillation) and self-refinement.
00:00 Introduction and inference-time scaling recap
05:02 Loading the pretrained LLM
08:00 Comparing and scoring model answers
10:18 Building a rule-based scorer
17:53 Token probabilities and sequence likelihood
26:47 Computing token probabilities in PyTorch
30:12 Token indexing and shifted targets
37:27 Log probabilities and numerical stability
45:57 Scoring answers with average log probabilities
56:24 How self-refinement works
59:07 Generating critiques and revised answers
1:01:00 Implementing the self-refinement loop
1:05:57 MATH-500 evaluation results
1:07:35 Takeaways and next steps
This video is larger than Cloudflare's 512 MB cache, so it can't be played through. More donations are needed to cover a larger cache. Donate
Agent-Editing World Model: Rethinking World Modeling for LLM Agents
huggingface.co/papers/2609.2…
Maximem Synap: The fastest, most accurate memory layer for AI agents by @gaurav_ships producthunt.com/products/max…
Opaline: Team-wide, message-level analytics for Claude Code and Codex by @evrendom producthunt.com/products/rud…
NOAN: Your Superhuman Business Partner by @imfilp producthunt.com/products/noa…
The Past Frames the Future: Memory for Autoregressive Video Generation
huggingface.co/papers/2609.2…
Jev State: Turn AI conversations into tests and runnable code by @yotta_mind producthunt.com/products/jev…
Koreshield: Security and evidence for AI support agents by @1cbyc producthunt.com/products/kor…
RankControl: Get Cited by ChatGPT & Ranked on Google by @WoCStreet producthunt.com/products/ran…
GBrain: Plug your own memory, tools, and skills into any AI by @bradgessler and @garrytan producthunt.com/products/gbr…
GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation
huggingface.co/papers/2609.2…
Hola AI: An AI voicemail assistant that answers calls when you can’t. by @thepmfguy producthunt.com/products/hol…
WeWeb.io: The best AI app builder for non-coders with big ideas by @joycekettering, @raphgoldz, @, and more producthunt.com/products/wew…
PixelCrew: Production-ready design from a crew of AI agents by @Askyruler and @antonpaisov producthunt.com/products/pix…
WZRD: AI-native documents, slides, forms and sheets that talk back by @alimo_ai and @eyemohd producthunt.com/products/wzr…
Anomalo: Your data is always talking. Don't miss what it's saying. by @eshmu producthunt.com/products/ano…