Hunee @Hunee

Ninja. Regular geek... Kitchen (office) Joined May 2007

Tweets

19K
Followers

529
Following

257
Likes

269

Hunee @Hunee

a week ago

and so my little project has to go back using Opus 4.8 because Fable 5 is not available for me. :(

0 0 0 32 0

View Details

Sam Witteveen @Sam_Witteveen

a month ago

Full breakdown on YT 👇 youtube.com/watch?v=0zHNyG…

0 2 4 316 0

View Details

Most agent frameworks put safety rules in the system prompt and hope the model follows them. Wrong layer — one injection and they're gone. OpenShell enforces policy via a YAML file. A supervisor sits outside the agent and controls: 🌐 network — default deny 📁 filesystem — sandboxed workspace 🧠 inference — managed endpoint 🔑 credentials — injected at runtime Compromise the agent, the policies still hold.

1 1 3 297 0

View Details

Sam Witteveen @Sam_Witteveen

a month ago

OpenShell is key for Agents - NemoClaw is a blueprint: - Harness (swappable) - Model (swappable) - Runtime — always OpenShell The runtime is the part doing the security work. Swap everything else, that stays. That's the piece worth understanding. DeepAgents can be that harness! @NVIDIAAI @NVIDIAAIDev @LangChain

1 2 8 533 3

View Details

Sam Witteveen @Sam_Witteveen

2 months ago

NVIDIA released Nemotron 3 Nano Omni. Open with a full paper of recipes for their data and SFT. One Open model. Text, images, video, audio — all in one. Mamba-MoE backbone, Vision encoder, Parakeet audio encoder.

1 2 7 573 2

View Details

Sam Witteveen @Sam_Witteveen

2 months ago

7 things every team needs locked down before shipping a multi-user agent to production. Rogue agents can run up a $10K bill overnight. Agents can hallucinate for 100s of users. Most builders don't notice. Here's what you should be monitoring. TL;DR 📋 🔀 Model control - unified layer between your code and models, swap providers without breaking your stack 📝 Prompt registry - version your prompts, keep them out of the codebase, treat them like a second tier of code 🛡️ Guardrails - pre and post LLM, pre and post tool calls. Handle PII, PHI, prompt injection, and output filtering 💰 Budget limiting - hard caps per model and project. The cloud providers won't do this for you 🔧 Tools & MCPs - central auth, granular permissions, test every one 📊 Monitoring & Tracing - trace every request, response, error, and latency spike ✅ Evals - before go-live and after. Catch the silent failures before your users do

1 5 14 588 5

View Details

Sam Witteveen @Sam_Witteveen

3 months ago

𝐆𝐞𝐦𝐦𝐚 𝟒 has landed with an 𝐀𝐩𝐚𝐜𝐡𝐞 𝟐.𝟎 𝐥𝐢𝐜𝐞𝐧𝐬𝐞. That alone would be news. But what's inside these models makes it a bigger deal. 𝗧𝗟;𝗗𝗥 📝 🏗️ 4 models across 2 tiers — 𝐖𝐨𝐫𝐤𝐬𝐭𝐚𝐭𝐢𝐨𝐧 (𝟐𝟳𝗕𝗔𝟰𝗕 𝗠𝗢𝗘, 𝟯𝟭𝗕 𝗱𝗲𝗻𝘀𝗲) and 𝗘𝗱𝗴𝗲 (𝗘𝟮𝗕, 𝗘𝟰𝗕) 🧠 128 experts in the MOE, only 3.8B active per token 🔊 Native audio support on Edge models with a 50% smaller audio encoder than Gemma 3N(681M → 305M params) 👁️ New vision encoder with variable aspect ratios and resolutions — much better OCR and document understanding 🤔 Built-in chain-of-thought 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 𝗮𝗰𝗿𝗼𝘀𝘀 𝘁𝗲𝘅𝘁, 𝗶𝗺𝗮𝗴𝗲𝘀, 𝗮𝗻𝗱 𝗮𝘂𝗱𝗶𝗼 🛠️ 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗰𝗮𝗹𝗹𝗶𝗻𝗴 𝗯𝗮𝗸𝗲𝗱 𝗶𝗻 from architecture level, not prompt-coaxed 📏 𝟭𝟮𝟴𝗞 𝗰𝗼𝗻𝘁𝗲𝘅𝘁 on edge, 𝟮𝟱𝟲𝗞 on workstation models 🌐 140 languages in pre-training, 35 in instruction tuning

1 3 16 682 1

View Details

VentureBeat @VentureBeat

4 months ago

Google Gemini 3.1 Pro first impressions: a 'Deep Think Mini' with adjustable reasoning on demand venturebeat.com/ai/google-gemi…

0 4 5 2K 3

View Details

Matt Marshall @mmarshall

4 months ago

@VentureBeat @Sam_Witteveen went deep on why this acquisition is the official obituary for the ChatGPT era. Read the breakdown on @VentureBeat here: venturebeat.com/technology/ope…

1 3 4 444 1

View Details

VentureBeat @VentureBeat

4 months ago

LangChain's CEO refused to let employees install OpenClaw on company laptops — then called it one of the most important agent projects in years. The paradox explains everything about where enterprise AI is heading. venturebeat.com/technology/ope…

8 4 14 5K 6

View Details

John Piazza @John_PiazzaIV

4 months ago

It's true: the web is not built for AI agents. But that's changing. @Sam_Witteveen summarizes the emerging WebMCP toolset from @googlechrome, @Microsoft, et al. Web browsing is expensive and inconsistently effective for agents. WebMCP changes that. venturebeat.com/infrastructure…

1 2 7 644 4

View Details

Hunee @Hunee

5 months ago

Big WIN for the open-source Community! Watch the full video @Eigent_AI @Sam_Witteveen

Sam Witteveen @Sam_Witteveen

5 months ago

Watch the Full video youtu.be/-UoxWCsqIa0

0 0 2 334 0

0 0 2 43 0

View Details

Sam Witteveen @Sam_Witteveen

5 months ago

Anthropic launched Cowork last week. One competing startup's response? Pivot their product and open-source EVERYTHING. The tweet announcing it hit 1.7M views. Here's the @Eigent_AI story 🧵

Guohao Li 🐫 @guohao_li

5 months ago

Anthropic Claude Cowork just killed our startup product 😅 So we did the most rational thing: open-sourced it. Meet Eigent 👉 github.com/eigent-ai/eige…

258 617 8K 1.8M 6K

1 5 15 4K 5

View Details

Sam Witteveen @Sam_Witteveen

5 months ago

Apple going with fine tuning of Gemini (Flash? / Flash Lite?) is probably smarter for them than jumping in to a fully "make your own models" strategy at this point in the game.

News from Google @NewsFromGoogle

5 months ago

Joint Statement: Apple and Google have entered into a multi-year collaboration under which the next generation of Apple Foundation Models will be based on Google's Gemini models and cloud technology. These models will help power future Apple Intelligence features, including a

2K 6K 52K 11.0M 6K

0 4 8 654 0

View Details

Sam Witteveen @Sam_Witteveen

5 months ago

🧵 Google just dropped UCP (Universal Commerce Protocol) at NRF with @sundarpichai himself announcing it 🛒 This is Google's play for agentic commerce - and it's built with Shopify, Etsy, Target, Walmart & Wayfair. Here's what you need to know 👇

1 3 9 492 3

View Details

Sam Witteveen @Sam_Witteveen

6 months ago

Happy New Year from Qwen!

@

57 years ago

0 0 0 1.7M 0

0 2 4 414 1

View Details

Sam Witteveen @Sam_Witteveen

6 months ago

Gemini 3 Flash is here! Google just dropped what could become your new daily driver model for AI applications. ⚡

3 3 24 4K 1

View Details

Sam Witteveen @Sam_Witteveen

6 months ago

Anti-Vibe Coding = Pair Programming with Top Models. Good to see Junie by @jetbrains focus on this style of coding agent for their IDEs 🧵

1 2 7 525 1

View Details

Google Cloud Tech @GoogleCloudTech

7 months ago

@TRJ_0751 @alarcon7a @mweinbach @tabGeeks Sam Witteveen tests the new "vibe coding" capabilities in AI Studio—from cloning classic games to building slick news sites. See what Gemini 3 built in a single shot 👇 youtube.com/watch?v=PFyccJ…

1 2 14 7K 4

View Details

Sam Witteveen @Sam_Witteveen

8 months ago

Kimi K2 Thinking by @Kimi_Moonshot just dropped and it's very impressive. This model is now beating OpenAI, Anthropic, and Google on major benchmarks. The LLM landscape just shifted dramatically.