←
Charly Wargnier ♨️36,414 posts
Charly Wargnier ♨️ avatar

Charly Wargnier ♨️

@DataChaz

Deep in Agents, LLMs and automation · Open source at ♥ · Sharing AI gems daily to help you work smarter · Ex-data scientist @Snowflake | @Streamlit ·

◷ Joined January 2009▧ 9,801 Photos and videos
Charly Wargnier ♨️ avatar
Charly Wargnier ♨️@DataChaz·1h

Most AI image tools break the moment you need a full campaign set. the new @Kling_ai Image 4.0 doesn't. one prompt, and you get an entire set, with the same color tone, design language and style from the hero visual to the detail pages quick dive in 🧵↓

Post attachmentPost attachmentPost attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·4h

this is a must-bookmark. Your next startup could start with this list. Every idea the @ycombinator team shared recently, curated here:

Post attachmentPost attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·2h

.@sama after watching old GPT-Sol crushing the new one

0:11

GPT-6.2 Astra is flat with GPT-6 Astra, and GPT-6.2 Sol loses 3 points relative to GPT-6.1 Sol in the Artificial Analysis Coding Agent Index (Codex harness at max effort).

Post attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·3h

THIS IS HUGE Google researchers just built an AI that learns while it “sleeps.” It updates its weights, stores new knowledge and practises on its own generated data. That tackles a stubborn problem. Training AI on new information can damage what it already knows. Bigger context windows help during a conversation, but they don’t automatically turn that information into lasting knowledge. Google Research’s Sleep framework splits learning into two phases: > Wake: Process inputs and collect short-term memories. > Sleep: Consolidate those memories into long-term parameters while preserving earlier knowledge. Then comes “dreaming.” The model generates synthetic examples, selects useful ones and trains on them to sharpen its capabilities. The reported results: • Near-perfect scaling on BABILong tests up to 10 million tokens. • Near-zero performance loss in sequential translation learning. • 79.2% on AIME 2024 and 69% on AIME 2025. “Sleep” is a training process inspired by biology. But the idea behind it is compelling. We keep giving AI more room to hold information. This research asks what happens when we also give it time to process that information. The next leap could happen after the conversation ends :)

Post attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·3h

freelancers realizing they have till midnight to land clients before GTA VI takes everyone offline for a month:

0:12
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·4h

the perks you can get with a student ID are just ridiculous.. if you're into tech, do yourself a favour and check them out:

Post attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·5h

Okay, this is so cool @nvidia has just dropped a 4GB model designed for local voice assistants. Nemotron 3.5 Bolt E4B combines conversational AI with multi-turn tool use. The focus? Respond quickly, use less memory, and leave space for the application running alongside it. think voice interfaces, interactive characters, and embedded assistants. Commercial and non-commercial use supported.

Post attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·16h

Anthropic valuation: $1.4 trillion Samsung valuation: $935 billion Anthropic 2025 revenue: $4.6 billion Samsung 2025 revenue: $236 billion

Post attachment
Charly Wargnier ♨️ avatar
Charly Wargnier ♨️@DataChaz·1h

and the text? well, it's clean! Big, bold branding stays sharp even on a dynamic shot, and the small supporting lines stay readable. Way less cleanup when I need impact and a clear message. Genuinely practical for the visual work I do around agents and demos.

Post attachment
Charly Wargnier ♨️ avatar
Charly Wargnier ♨️@DataChaz·1h

Zoom in and the faces still hold up. Skin texture, lighting, tiny surface details: all natural enough to crop or blow up for decks and product assets. Native 8K makes that actually practical.

Post attachment
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·4h

you don't need to learn agent frameworks this weekend. you just need this prompt and Claude Fable 5.2. thank me later.

Post attachment
Charly Wargnier ♨️ avatar
Charly Wargnier ♨️@DataChaz·3h

the most important calls still happen on the phone. ... and they're the ones nobody takes notes on! @bluedot_ai just fixed that: > call from the app, phone or web > recording, transcript and AI notes when you hang up > synced to your CRM, Notion, Claude, ChatGPT love it.

New in Bluedot: phone call recording. Call anyone from the app. Get the notes when you hang up.

0:21
Charly Wargnier ♨️ avatar
⟳ Charly Wargnier ♨️ retweeted
Charly Wargnier ♨️@DataChaz·17h

Australia is officially entering the AI race. @AgentausTD just threw a 220B model on the barbie

Post attachment

We built Australia’s own AI model, and today it’s open to the world. Agentaus 1 (@AgentausTD): 220B parameters. 84.4% on MMLU-Pro. Ranked 43rd in the world. It runs on Australian servers. Nothing ever leaves Australia, and nothing you type is used to train AI. Delete a chat and it’s gone for good. Most of the world’s AI answers to somebody else. This one answers only to you. Free to start: agentaus.com.au

1:54