Charis Tsevis
banner
tsevis.bsky.social
Charis Tsevis
@tsevis.bsky.social
Visual designer living on the sunny side of the planet.
http://www.tsevis.com
From sci-paper analysis to sophisticated software. With basic skills, I used ClaudeAI as a partner to build it. My story:
tsevis.com/how-i-learne...
#GenAI #ClaudeCode
September 1, 2025 at 8:36 PM
OpenAI keeps delaying its first ‘open’ model, saying it needs to be made bulletproof. After watching this video, you’ll see why safety is a massive problem with today’s AI. Sure, you could read about jailbreaks in some dry research paper.
#AISafety #AIHack
August 17, 2025 at 5:24 PM
Why is it my new go-to? While CLIP (from OpenAI or others) requires text captions and specialist models that overfit to niche data, DINOv3 operates like pure visual intuition. #SSLforImages
August 15, 2025 at 3:23 PM
Forget begging for labeled data or wrestling with five different models. This model learns purely from raw pixels, 1.7 billion of them, and provides a single, frozen backbone that handles tasks such as depth mapping, segmentation, and material estimation out of the box.
#SSL #AI
August 15, 2025 at 3:23 PM
Computer Vision is probably my favorite discipline of IT.
If you are building apps that enable your computer to see (style transfer, object removal, smart masking, visual effects etc.) with DINOv3 has quietly changed the game.
#ComputerVision
August 15, 2025 at 3:23 PM
Also use weighted prompts based on relevance.
Works across 6 tasks. No fine-tuning. No extra data.
Interesting step in using models more intelligently, not just bigger or trained more.
arxiv.org/html/2508.09...
#ComputerVision #ML
August 14, 2025 at 8:50 AM
A new paper shows how Stable Diffusion can do visual in-context learning without any training.
They change how attention works during inference to let the model compare a query image with example prompts.
August 14, 2025 at 8:50 AM
Made with custom algorithms + 15 years of African textile research. 4th edition in my series.
#DigitalArt #AfricanCulture
August 13, 2025 at 10:32 AM
Two parts: 5 imaginary women (archetypal spirits) + 5 real icons (Tyla, Siya Kolisi, Thuli Madonsela). Each uses 1000+ textile patterns. #AlgorithmicArt #ComputationalArt
August 13, 2025 at 10:32 AM
Sometimes you need to trace backward to understand where you're going. The full story surprised even me.
You can read the piece on my blog: tsevis.com/when-machine...
#DigitalArt #DigitalComplexity
August 5, 2025 at 10:55 AM
The things he taught me in his workshop now mix with my computer coding. Hands-on meets digital. #CreativeCode #CodeArt #DigitalMosaic
August 5, 2025 at 10:55 AM
Dad's lesson: "Find poetry in the machine." Every constraint became creativity. Every limitation, liberation.
August 5, 2025 at 10:55 AM
The path: 1898 typewriter art to early computer experts at Bell Labs to groups making art with keyboard symbols to my computer programs today. Each era builds on the last. #AsciiArt #AnsiArt #8bitbeauty #DigitalMosaic #Photomosaic
August 5, 2025 at 10:55 AM
Spent countless hours writing about how my dad shaped my digital art journey - from his inventive spirit in 1970s Greece to Fortune 500 work. #IT #ASCIIArt #MachineArt #Typography
August 5, 2025 at 10:55 AM
The devastating truth about current AI models: They have no real understanding. There is no intelligence in them.
New Harvard/MIT research trained models on planetary orbits. They perfectly predicted planetary motion but completely failed when asked to apply basic physics (F=ma) to new scenarios.
August 3, 2025 at 5:42 PM
Most revealing moments: Named Tesla at Disneyland, "gave away too much equity" when Musk invested, had to sue Tesla just to get credit in the first Roadster, and thinks current self-driving promises are overhyped. "I don't want a self-driving car."
August 3, 2025 at 9:23 AM
The Tesla founder you've never heard of tells his side of the story. Martin Eberhard co-founded Tesla but got overshadowed by history. His perspective on the early days is eye-opening and changes everything you think you know about Tesla's origin. #Tesla #elonmusk #ev
August 3, 2025 at 9:23 AM
Must-watch interview: Dario Amodei gets candid about AI, Anthropic, and being called a "doomer":
Most powerful moments:
His father died from a disease that became 95% curable just 3-4 years later - this personal loss drives his urgency around AI's potential to accelerate medical breakthroughs.
#AI
August 2, 2025 at 1:30 PM
GLM-4.5 by Z.ai is here!
• Top-tier benchmark scores across reasoning & coding
• Elite coding skills → just one step below Claude.ai
• Advanced multilingual capabilities
• Apache 2.0 licensed • Open-source
• Optimized for complex problem-solving #opensourceai #localllm #GenAi
z.ai/blog/glm-4.5...
August 2, 2025 at 10:46 AM
Just watched an amazing Claude Code best practices talk by Cal Rueb from Anthropic. His top tips:
Use claude . md files for persistent context across sessions
Ask Claude to plan before coding (saves tons of time)
Run multiple Claude instances in parallel for complex projects
July 31, 2025 at 8:05 PM
Indian truck art = moving masterpieces
• Symbols of love, faith and humor
• Hand-painted on trucks across the country
• Bold colors, poetic slogans, divine icons
• A visual feast rolling on every highwayWorla
Deeply amazed by the creativity.
enrouteindianhistory.com/why-is-truck...
July 28, 2025 at 10:27 AM
Qwen3 Coder dropped!
• 0.5B to 72B parameters
• Open source (Apache 2.0 license)
• Trained for code generation, completion, and understanding
• Supports multiple programming languages
• Chat-style coding assistant
github.com/QwenLM/Qwen3...
July 24, 2025 at 9:33 PM
My new piece explores the journey from admiration to disappointment, and what happens when ego outpaces vision. #ArtisticResponsibility #DigitalMosaic
June 29, 2025 at 10:13 AM
Did I help build the myth?
As the artist who created Elon Musk's TIME Person of the Year portrait, I'm grappling with a hard question about artistic responsibility. #ElonMusk #Storytelling #CreativeEthics
June 29, 2025 at 10:13 AM
GenAI's "Diffusion Effect" is like painting: broad strokes, then adding details. Text generation refines outputs iteratively, making conversations better and faster. Pretty neat, right?
Try it at @inceptionlabs.bsky.social :
chat.inceptionlabs.ai #LLM #GenAI #MercuryLLM
June 29, 2025 at 8:24 AM