AgentSkills

Tech enthusiast tracking the cutting edge of AI, robotics, Python open source, Gaussian splatting, and medical imaging.

[AgentSkills @CodeTrendr]

Strata 0.1.42 → 0.1.43 on RTX 4090 + Qwen3.8 Flash-Next:

• 21–37% faster decoding
• 137–157 tokens/s across 4K–380K prompts
• 380K retrieval: 8/8 facts recovered

Prefill was unchanged. Same hardware/settings, 3 speed runs per size.

#LocalLLM #Strata

[Hasan Toor @hasantoxr]

🚨BREAKING: Someone just solved Claude Code's biggest problem.

It's called Claude-Mem and it gives Claude persistent memory across sessions.

- You can use up to 95% fewer tokens each time.
- Make 20 times more tool calls before reaching limits.

100% Opensource.

[Confidence @megaconfidence]

Stop buying $600 mac minis!

Run @openclaw for $5 on Cloudflare Workers

https://t.co/9b7z4UfLqz

[Google Gemini @GeminiApp]

Introducing Agentic Vision, a new capability in Gemini 3 Flash.

Agentic Vision makes Gemini even better at analyzing complex images, enabling it to more accurately and consistently read fine details, like serial numbers or text on a complex diagram.

See what it can do. 🧵

[ollama @ollama]

🎉🎉🎉 Kimi K2.5 is on Ollama's cloud

ollama run kimi-k2.5:cloud

You can connect it to Claude Code, Codex, OpenCode, Clawdbot, and Droid via ollama launch!

ollama launch claude --model kimi-k2.5:cloud

[AgentSkills @CodeTrendr]

@rsasaki0109 Wow, I think someone are waiting it for years! Great job

[NVIDIA AI Developer @NVIDIAAIDev]

🏗️ Most 3D reconstruction papers evaluate novel views by applying an optimal color correction using the ground-truth image. That's circular and unrealistic.

PPISP directly predicts camera parameters from the rendered radiance alone, using an ISP controller trained on input views.

✨Result: metrics that don't require access to target images, aligning evaluation with real-world deployment.

➡️ https://t.co/ox003dGf1Y

[Dhanian 🗯️ @e_opore]

Full Stack Folder Structure.

Grab the Modern Full Stack Latest Edition Handbook: https://t.co/cQjcyhz9KR

[Hyper3D by Deemos @DeemosTech]

💥🍌3D Nano Banana just dropped!
✨We just launched #Rodin Gen-2 "Edit", upload ANY model and edit like magic:

1️⃣ Smart Low-poly → artist-style topology
2️⃣ Local edits via prompt (Beta)
3️⃣ BANG to Parts
4️⃣+ more

#Hyper3D is now the FIRST true #3D GenAI editing platform! 🚀

[Akshay 🚀 @akshay_pachaar]

Big moment for Postgres!

Search has always been Postgres' weak spot, and everyone just accepted it.

If you needed a real relevance-ranked keyword search, the default answer was to spin up Elasticsearch or add Algolia and deal with the data sync headaches forever.

The problem isn't that Postgres can't do text search. It can.

But the built-in `ts_rank` function uses a basic term frequency algorithm that doesn't come close to what modern search engines deliver.

So teams end up:

- Running a separate Elasticsearch cluster just for search
- Building sync pipelines that inevitably drift out of consistency
- Paying for managed search services that charge per query
- Accepting mediocre search relevance because "good enough" ships faster

But this is actually a solvable problem.

You can realistically bring industry-standard search ranking directly into Postgres, which eliminates the need for external infra entirely.

This exact solution is now available with the newly open-sourced pg_textsearch by @TigerDatabase, a Postgres extension that brings true BM25 relevance ranking into the database.

BM25 is the algorithm behind Elasticsearch, Lucene, and most modern search engines. Now it runs natively in Postgres.

Here's what pg_textsearch enables:

- True BM25 ranking with configurable parameters (the same algorithm powering production search systems)
- Simple SQL syntax: `ORDER BY content <@> 'search terms'`
- Works with Postgres text search configurations for multiple languages
- Pairs naturally with pgvector for hybrid keyword + semantic search

That last point matters a lot for RAG apps. The video below shows this in action, and I worked with the team to put this together.

You can now do hybrid retrieval (combining keyword matching with vector similarity) in a single database, without stitching together multiple systems.

The syntax is clean enough that you can add relevance-ranked search to existing queries in minutes.

pg_textsearch is fully open-source under the PostgreSQL license.

You can find a link to their GitHub repo in the next tweet.

[ERNIE for Developers @ErnieforDevs]

🚀 ERNIE 5.0 is officially live!
Supporting text, image, audio, and video inputs & outputs, ERNIE 5.0 ranks in the top tier across multiple authoritative benchmarks.

✅ 2.4T total parameters
✅ Stronger language & multimodal understanding
✅ Image & video generation on par with specialized domain models

👉Try it now:
Chat: https://t.co/jNwEbwrczU
Playground: https://t.co/UAYkWfflIK
API: https://t.co/6kxaUJDvdl

[Tom Krcha @tomkrcha]

Excited to launch Pencil
INFINITE DESIGN CANVAS for Claude Code

&gt; Superfast WebGL canvas, fully editable, running parallel design agents
&gt; Runs locally with Claude Code → turn designs into code
&gt; Design files live in your git repo → Open json-based .pen format

[Stitch by Google @stitchbygoogle]

Developer Week Day 2. 🛠️

We just gave your Coding Agent a Design Degree🎓🎨

Introducing the Stitch MCP Server. 🔌

You can now pipe Stitch designs directly into your favorite tools like Antigravity.

* Generate new screens without leaving your IDE
* Fetch the code from any design
* Inject context: Give your agent full visual awareness

Docs and more information 👇

[World Labs @theworldlabs]

The World API is live.

Generate persistent, explorable 3D worlds from text, images, and video.

Integrate them directly into your products.

[Unity for Games @unitygames]

The @Walmart Immersive Commerce SDK is now available! Bring real-world products into gameplay by integrating Walmart items or your own marketplace merchandise directly into your experience without breaking immersion.

Start building today: https://t.co/a5nFdXTcy8

[MeshyAI @MeshyAI]

Image to 3D, now in ANY pose! 🤸‍♂️

Simply upload your image + a pose reference photo. Meshy generates the model to match it perfectly.

No complex rigging needed. Create 3D characters your way.

[Charly Wargnier ♨️ @DataChaz]

NVIDIA just removed one of the biggest friction points in Voice AI.

PersonaPlex-7B is an open-source, full-duplex conversational model.

Free, open source (MIT), with open model weights on @huggingface 🤗

Links to repo and weights in 🧵↓

The traditional ASR → LLM → TTS pipeline forces rigid turn-taking.
It’s efficient, but it never feels natural.

PersonaPlex-7B changes that.

This @nvidia model can listen and speak at the same time.

It runs directly on continuous audio tokens with a dual-stream transformer, generating text and audio in parallel instead of passing control between components.

That unlocks:
→ instant back-channel responses
→ interruptions that feel human
→ real conversational rhythm

Persona control is fully zero-shot!

If you’re building low-latency assistants or support agents, this is a big step forward 🔥

[Said Taghadouini @staghado]

🚀 LightOnOCR-2-1B 🦉 is out, a major update to LightOnOCR.
1B parameters, end-to-end multilingual OCR, and it beats models 9× larger on OlmOCR-Bench while being much faster.
PDF/page in, clean ordered Markdown out, with optional image localization (bbox variants).

[Yawar Siddiqui @yawarnihal]

Introducing ShapeR, a method for robust conditional 3D shape generation from casually captured sequences.

ShapeR leverages a rectified flow transformer conditioned on per-object multimodal data to turn casual image sequences into full metric scene reconstructions.

Project Page: https://t.co/ffH2zVKd8c
Paper: https://t.co/82yWRrXnA1

Links to code and huggingface below ⬇️

[AgentSkills @CodeTrendr]

@averycode Even though we don’t know each other, I send you my best wishes for your brave decision.