Ahmed Fessi
@ahmedfe.bsky.social
Nested Learning lets models adapt without forgetting, saving retraining time, but adds complexity and risk of bias propagation. #AI #MachineLearning #ContinualLearning #DataScience #DigitalTransformation
research.google/blog/introdu...
research.google/blog/introdu...
Introducing Nested Learning: A new ML paradigm for continual learning
We strive to create an environment conducive to many different types of research across many different time scales and levels of risk.
research.google
November 12, 2025 at 7:35 AM
Nested Learning lets models adapt without forgetting, saving retraining time, but adds complexity and risk of bias propagation. #AI #MachineLearning #ContinualLearning #DataScience #DigitalTransformation
research.google/blog/introdu...
research.google/blog/introdu...
EuroLLM opens opportunities for European developers and researchers to build models that reflect local rules and languages.
#AI #ML #DataPrivacy #DigitalEurope #ResponsibleAI
eurollm.io
#AI #ML #DataPrivacy #DigitalEurope #ResponsibleAI
eurollm.io
November 6, 2025 at 1:08 PM
EuroLLM opens opportunities for European developers and researchers to build models that reflect local rules and languages.
#AI #ML #DataPrivacy #DigitalEurope #ResponsibleAI
eurollm.io
#AI #ML #DataPrivacy #DigitalEurope #ResponsibleAI
eurollm.io
AI (Perplexity) is making patent insight instant.
#AI #Innovation #DigitalTransformation
www.perplexity.ai/hub/blog/int...
#AI #Innovation #DigitalTransformation
www.perplexity.ai/hub/blog/int...
November 6, 2025 at 10:40 AM
AI (Perplexity) is making patent insight instant.
#AI #Innovation #DigitalTransformation
www.perplexity.ai/hub/blog/int...
#AI #Innovation #DigitalTransformation
www.perplexity.ai/hub/blog/int...
Anthropic’s new paper hints at something intriguing: models that can sometimes detect their own injected thoughts.
#AI #LLM #Anthropic #AIsafety
www.anthropic.com/research/int...
#AI #LLM #Anthropic #AIsafety
www.anthropic.com/research/int...
November 6, 2025 at 6:57 AM
Anthropic’s new paper hints at something intriguing: models that can sometimes detect their own injected thoughts.
#AI #LLM #Anthropic #AIsafety
www.anthropic.com/research/int...
#AI #LLM #Anthropic #AIsafety
www.anthropic.com/research/int...
Pipelex - a simple and practical way to start building basic agents
#AI #DataEngineering #MLOps #Automation
github.com/Pipelex/pipe...
#AI #DataEngineering #MLOps #Automation
github.com/Pipelex/pipe...
November 5, 2025 at 5:06 PM
Pipelex - a simple and practical way to start building basic agents
#AI #DataEngineering #MLOps #Automation
github.com/Pipelex/pipe...
#AI #DataEngineering #MLOps #Automation
github.com/Pipelex/pipe...
AI tools are now learning from the small things we write for humans like code comments.
#AI #DataEthics #DigitalTransformation
cryptography.dog/blog/AI-scra...
#AI #DataEthics #DigitalTransformation
cryptography.dog/blog/AI-scra...
November 5, 2025 at 1:13 PM
AI tools are now learning from the small things we write for humans like code comments.
#AI #DataEthics #DigitalTransformation
cryptography.dog/blog/AI-scra...
#AI #DataEthics #DigitalTransformation
cryptography.dog/blog/AI-scra...
Microsoft 365 Copilot can now run tasks on a virtual computer.
#Microsoft365 #Copilot #AI #DataSecurity #Tech
techcommunity.microsoft.com/blog/microso...
#Microsoft365 #Copilot #AI #DataSecurity #Tech
techcommunity.microsoft.com/blog/microso...
November 5, 2025 at 7:35 AM
Microsoft 365 Copilot can now run tasks on a virtual computer.
#Microsoft365 #Copilot #AI #DataSecurity #Tech
techcommunity.microsoft.com/blog/microso...
#Microsoft365 #Copilot #AI #DataSecurity #Tech
techcommunity.microsoft.com/blog/microso...
Interesting addition to AI / OpenSource / Agents : teaching agents to plan, reason, and collaborate - CheckTongyi Deep Research.
#AI #Transformation #TechInsight
tongyi-agent.github.io/blog/introdu...
#AI #Transformation #TechInsight
tongyi-agent.github.io/blog/introdu...
November 4, 2025 at 5:28 PM
Interesting addition to AI / OpenSource / Agents : teaching agents to plan, reason, and collaborate - CheckTongyi Deep Research.
#AI #Transformation #TechInsight
tongyi-agent.github.io/blog/introdu...
#AI #Transformation #TechInsight
tongyi-agent.github.io/blog/introdu...
November 3, 2025 at 3:31 PM
Claude Haiku 4.5 can now handle images and documents, not just text. The direction is toward AI that works instantly and reliably in everyday tools.
#AI #Claude #Anthropic
www.anthropic.com/news/claude-...
#AI #Claude #Anthropic
www.anthropic.com/news/claude-...
Introducing Claude Haiku 4.5
Claude Haiku 4.5, our latest small model, is available today to all users.
www.anthropic.com
October 24, 2025 at 9:04 AM
Claude Haiku 4.5 can now handle images and documents, not just text. The direction is toward AI that works instantly and reliably in everyday tools.
#AI #Claude #Anthropic
www.anthropic.com/news/claude-...
#AI #Claude #Anthropic
www.anthropic.com/news/claude-...
TPTL #10 is live!
Inside this issue:
• GPT Atlas, OpenAI’s new browser
• Google’s Willow quantum chip
• Sora and the deepfake dilemma
• Why hallucinations are inevitable in LLMs
• The world’s first AI Minister
#AI #Quantum #Innovation #TPTL
the-pragmatic-tech-leader.kit.com/posts/tptl-1...
Inside this issue:
• GPT Atlas, OpenAI’s new browser
• Google’s Willow quantum chip
• Sora and the deepfake dilemma
• Why hallucinations are inevitable in LLMs
• The world’s first AI Minister
#AI #Quantum #Innovation #TPTL
the-pragmatic-tech-leader.kit.com/posts/tptl-1...
TPTL #10 - Your Monthly Insights on Tech, AI & LLMs: GPT Atlas, Willow Quantum Chip, SORA, and more
OpenAI’s Atlas browser, Google’s quantum leap, Sora’s legal storm, AI hallucinations, and Albania’s AI minister, all in TPTL #10.
the-pragmatic-tech-leader.kit.com
October 23, 2025 at 3:22 PM
TPTL #10 is live!
Inside this issue:
• GPT Atlas, OpenAI’s new browser
• Google’s Willow quantum chip
• Sora and the deepfake dilemma
• Why hallucinations are inevitable in LLMs
• The world’s first AI Minister
#AI #Quantum #Innovation #TPTL
the-pragmatic-tech-leader.kit.com/posts/tptl-1...
Inside this issue:
• GPT Atlas, OpenAI’s new browser
• Google’s Willow quantum chip
• Sora and the deepfake dilemma
• Why hallucinations are inevitable in LLMs
• The world’s first AI Minister
#AI #Quantum #Innovation #TPTL
the-pragmatic-tech-leader.kit.com/posts/tptl-1...
EMM1 evaluates how AI understands images and text together. It highlights where models excel and where they fall short, helping build more reliable multimodal systems.
#AI #Data #VisionLanguage
encord.com/multimodal-d...
#AI #Data #VisionLanguage
encord.com/multimodal-d...
E-MM1 Dataset: The World's Largest Multimodal AI Dataset
The E-MM1 dataset is the world's largest multimodal AI dataset, with more than 100 million groups of data in five modalities to foster the development of models that fuse multiple modalities.
encord.com
October 23, 2025 at 12:08 PM
EMM1 evaluates how AI understands images and text together. It highlights where models excel and where they fall short, helping build more reliable multimodal systems.
#AI #Data #VisionLanguage
encord.com/multimodal-d...
#AI #Data #VisionLanguage
encord.com/multimodal-d...
ChatGPT now organizes memories with prioritization.
#AI #Data #DigitalTransformation
help.openai.com/en/articles/...
#AI #Data #DigitalTransformation
help.openai.com/en/articles/...
ChatGPT — Release Notes | OpenAI Help Center
A changelog of the latest updates for ChatGPT
help.openai.com
October 23, 2025 at 9:40 AM
ChatGPT now organizes memories with prioritization.
#AI #Data #DigitalTransformation
help.openai.com/en/articles/...
#AI #Data #DigitalTransformation
help.openai.com/en/articles/...
New research shows small AI models can reason surprisingly well when built recursively.
It’s a reminder that efficiency and smart design can rival size.
Less compute, more thinking. #AI #DeepLearning #Innovation
arxiv.org/html/2510.04...
It’s a reminder that efficiency and smart design can rival size.
Less compute, more thinking. #AI #DeepLearning #Innovation
arxiv.org/html/2510.04...
October 23, 2025 at 5:57 AM
New research shows small AI models can reason surprisingly well when built recursively.
It’s a reminder that efficiency and smart design can rival size.
Less compute, more thinking. #AI #DeepLearning #Innovation
arxiv.org/html/2510.04...
It’s a reminder that efficiency and smart design can rival size.
Less compute, more thinking. #AI #DeepLearning #Innovation
arxiv.org/html/2510.04...
OpenAI just released Atlas, a new Browser.
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
www.linkedin.com/feed/update/...
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
www.linkedin.com/feed/update/...
OpenAI just released Atlas, a new Browser.
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
But it also raises… | ...
OpenAI just released Atlas, a new Browser.
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
But it also raises dee...
www.linkedin.com
October 21, 2025 at 9:16 PM
OpenAI just released Atlas, a new Browser.
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
www.linkedin.com/feed/update/...
It’s hard not to see this as a move towards direct competition with the tech giants, especially Google, both in browsing and search.
www.linkedin.com/feed/update/...
OpenAI is working with Broadcom to make special AI chips. They'll use 10 gigawatts of processing power for things like ChatGPT and Sora. It's a move away from Nvidia, changing how AI tech is built from the ground up.
#OpenAI #AI
#OpenAI #AI
OpenAI is teaming up with Broadcom to create their own AI chips
OpenAI is collaborating with Broadcom to develop its own computer chips specifically for its AI data centers. It is one of several…
medium.com
October 21, 2025 at 6:35 AM
Linux refined how it handles interrupts, making systems more predictable and stable.
Small change, real impact for AI and data infrastructure.
#Linux #AIInfrastructure #Engineering
lwn.net/Articles/103...
Small change, real impact for AI and data infrastructure.
#Linux #AIInfrastructure #Engineering
lwn.net/Articles/103...
October 19, 2025 at 7:21 AM
Linux refined how it handles interrupts, making systems more predictable and stable.
Small change, real impact for AI and data infrastructure.
#Linux #AIInfrastructure #Engineering
lwn.net/Articles/103...
Small change, real impact for AI and data infrastructure.
#Linux #AIInfrastructure #Engineering
lwn.net/Articles/103...
Your data model quietly decides how your product grows.
When it’s clear, every feature aligns. When it’s messy, everything drifts.
#DataStrategy #ProductDesign #AI
notes.mtb.xyz/p/your-data-...
When it’s clear, every feature aligns. When it’s messy, everything drifts.
#DataStrategy #ProductDesign #AI
notes.mtb.xyz/p/your-data-...
October 19, 2025 at 6:51 AM
Your data model quietly decides how your product grows.
When it’s clear, every feature aligns. When it’s messy, everything drifts.
#DataStrategy #ProductDesign #AI
notes.mtb.xyz/p/your-data-...
When it’s clear, every feature aligns. When it’s messy, everything drifts.
#DataStrategy #ProductDesign #AI
notes.mtb.xyz/p/your-data-...
OpenAI is making its own AI chips with Broadcom. These chips, using 10 gigawatts of power, will help power ChatGPT and Sora. This move cuts back OpenAI's reliance on Nvidia and changes AI tech.
#OpenAI #AI #Sora
#OpenAI #AI #Sora
OpenAI is teaming up with Broadcom to create their own AI chips
OpenAI is collaborating with Broadcom to develop its own computer chips specifically for its AI data centers. It is one of several…
medium.com
October 17, 2025 at 5:07 PM
Apple’s M5 chip boosts on-device AI performance.
Faster, efficient, and ready for local inference.
#AI #Data #DigitalTransformation
www.apple.com/newsroom/202...
Faster, efficient, and ready for local inference.
#AI #Data #DigitalTransformation
www.apple.com/newsroom/202...
October 17, 2025 at 1:06 PM
Apple’s M5 chip boosts on-device AI performance.
Faster, efficient, and ready for local inference.
#AI #Data #DigitalTransformation
www.apple.com/newsroom/202...
Faster, efficient, and ready for local inference.
#AI #Data #DigitalTransformation
www.apple.com/newsroom/202...
Social engineering keeps getting smarter. Stay cautious.
#CyberSecurity #AI #DigitalTransformation
blog.daviddodda.com/how-i-almost...
#CyberSecurity #AI #DigitalTransformation
blog.daviddodda.com/how-i-almost...
October 17, 2025 at 9:04 AM
Social engineering keeps getting smarter. Stay cautious.
#CyberSecurity #AI #DigitalTransformation
blog.daviddodda.com/how-i-almost...
#CyberSecurity #AI #DigitalTransformation
blog.daviddodda.com/how-i-almost...
Karpathy’s NanoChat: A few hundred lines, full chatbot.
#AI #LLMs #OpenSource #Python #DigitalTransformation
Link:
#AI #LLMs #OpenSource #Python #DigitalTransformation
Link:
GitHub - karpathy/nanochat: The best ChatGPT that $100 can buy.
The best ChatGPT that $100 can buy. Contribute to karpathy/nanochat development by creating an account on GitHub.
github.com
October 14, 2025 at 12:32 PM
Karpathy’s NanoChat: A few hundred lines, full chatbot.
#AI #LLMs #OpenSource #Python #DigitalTransformation
Link:
#AI #LLMs #OpenSource #Python #DigitalTransformation
Link:
Even a few poisoned samples can quietly corrupt an AI model, no matter how large the dataset.
Data hygiene isn’t maintenance it’s security.
More data isn’t safer. Better data is.
#AI #DataSecurity #DigitalTransformation
www.linkedin.com/feed/update/...
Data hygiene isn’t maintenance it’s security.
More data isn’t safer. Better data is.
#AI #DataSecurity #DigitalTransformation
www.linkedin.com/feed/update/...
www.linkedin.com
October 14, 2025 at 6:35 AM
Even a few poisoned samples can quietly corrupt an AI model, no matter how large the dataset.
Data hygiene isn’t maintenance it’s security.
More data isn’t safer. Better data is.
#AI #DataSecurity #DigitalTransformation
www.linkedin.com/feed/update/...
Data hygiene isn’t maintenance it’s security.
More data isn’t safer. Better data is.
#AI #DataSecurity #DigitalTransformation
www.linkedin.com/feed/update/...
LLMs still feel off when coding they don’t copy-paste, they recreate from memory, and they rarely ask questions.
Fast, yes. But not thoughtful.
.
#AI #LLM #Coding #SoftwareDevelopment #DigitalTransformation #AItools #TechInsights
kix.dev/two-things-l...
Fast, yes. But not thoughtful.
.
#AI #LLM #Coding #SoftwareDevelopment #DigitalTransformation #AItools #TechInsights
kix.dev/two-things-l...
October 13, 2025 at 2:31 PM
LLMs still feel off when coding they don’t copy-paste, they recreate from memory, and they rarely ask questions.
Fast, yes. But not thoughtful.
.
#AI #LLM #Coding #SoftwareDevelopment #DigitalTransformation #AItools #TechInsights
kix.dev/two-things-l...
Fast, yes. But not thoughtful.
.
#AI #LLM #Coding #SoftwareDevelopment #DigitalTransformation #AItools #TechInsights
kix.dev/two-things-l...