Reda

Airing

[Reda @redaelhail]

@kasmiyouness1 I think google already proposed synthID for detecting AI generated video and images

[Gregor @bygregorr]

Everyone saw the Microsoft headline this week and read it as another billion dollar AI flex. The actual signal is way more useful than that.

Microsoft put 2.5 billion and 6000 engineers into one new group. Their job is not building models. It’s going inside companies and making AI deployments actually work in a real business.

Every big lab landed on the same realization around the same time. Selling access to a smart model doesn’t stick. The money is in the last mile, the unglamorous part where you sit with one company’s broken process and wire the model into it until it finishes the work with nobody watching.

And that last mile does not scale for them. Too specific, too manual, too small per customer. Which is exactly why it’s open to you.

Nobody at a foundation lab is going to out-work you on claims processing for physio clinics, or quote generation for HVAC installers, or whatever narrow thing you already understand better than an outsider ever could.

The models are a commodity now. The context around them isn’t.

[Reda @redaelhail]

@kasmiyouness1 Are the videos good quality?

[Youness KASMI @kasmiyouness1]

You can now get FREE .ai.studio subdomain at Google AI Studio. But be careful, if you are publishing a static website, make sure it is not using your Gemini API key. Otherwise, your visitors will consume your tokens. If it's necessary for your app, don't forget to put a spend cap

[Reda @redaelhail]

Comparison is different benchmarks

[Reda @redaelhail]

Muse Spark 1.1 AI model priced at $1.25 per million input tokens and $4.25 per million output tokens. Significantly undercutting OpenAI's GPT-5.5 ($5/$30) and Anthropic's Claude Opus 4.8 ($5/$25)

#AI #META #claude #openai

[Youness KASMI @kasmiyouness1]

Prompt engineering shouldn't be that hard. It's just giving clear instructions. If you are an AI builder, give your AI a role, context, tell it what not to do, and explain the desired outcome. Quality output starts with quality input. Stop wasting tokens on bad prompts.

[Reda @redaelhail]

Great insights from @daedalium

[Oussama Ammar @daedalium]

Live with Ouss | Blabla & Bonne Humeur https://t.co/Oy85snQs8H

[David Ondrej @DavidOndrej1]

im releasing all my agent skills to the public

this is hundreds of hours of trial & error

every single global .agents skill i have

go grab it. it's free.

[Reda @redaelhail]

@muslehNayesh ما مدى فعاليته ؟ @grok

[Google Gemma @googlegemma]

Meet Gemma 4 12B!

A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license.

Bridging the gap between edge efficiency and advanced reasoning. Here is what’s new with Gemma 4 12B: 👇

[Reda @redaelhail]

Arthur mensch, founder of MISTRAL AI talking about how important marketing and the ability to sell your product is important. It's an indication on how important it is to learn this skill rather than only focusing on technical aspects. https://t.co/PcpM75HmRs

[Joseph Carlson @joecarlsonshow]

I invested a total of $96,000 into Google, and it's now worth $222,000. Thank you, Sundar and team.

[Mohammed Daoudi @MIduoad]

Mrrakc is not a tourism app, and not a maps website! Mrrakc is an open Dataset for tourism in Morocco.

The idea started with my friends asking me for places' recommendations in Morocco, and each I had to send them long messages, audios, or saved maps links.

[Reda @redaelhail]

@GoldmanSachs Fast and intuitive jobs vs. slow and logical thinking jobs. AI is currently swallowing the former ones.

[Reda @redaelhail]

@Med_Roudan أمن و استقرار الأبناك مهم لأن نظامالبنك مبني على ثقة المواطن في البنك و إلا مشات الثقة كلشي يبغي يجبد فلوسو. الاقتصاد أيتوقف

[Aakash Gupta @aakashgupta]

Someone just poisoned the Python package that manages AI API keys for NASA, Netflix, Stripe, and NVIDIA.. 97 million downloads a month.. and a simple pip install was enough to steal everything on your machine.

The attacker picked the one package whose entire job is holding every AI credential in the organization in one place. OpenAI keys, Anthropic keys, Google keys, Amazon keys… all routed through one proxy. All compromised at once.

The poisoned version was published straight to PyPI.. no code on GitHub.. no release tag.. no review. Just a file that Python runs automatically on startup. You didn’t need to import it. You didn’t need to call it. The malware fired the second the package existed on your machine.

The attacker vibe coded it… the malware was so sloppy it crashed computers.. used so much RAM a developer noticed their machine dying and investigated. They found LiteLLM had been pulled in through a Cursor MCP plugin they didn’t even know they had.

That crash is the only reason thousands of companies aren’t fully exfiltrated right now. If the code had been cleaner nobody notices for weeks. Maybe months.

The attack chain is the part that gets worse every sentence.

TeamPCP compromised Trivy first. A security scanning tool. On March 19. LiteLLM used Trivy in its own CI pipeline… so the credentials stolen from the SECURITY product were used to hijack the AI product that holds all your other credentials.

Then they hit GitHub Actions. Then Docker Hub. Then npm. Then Open VSX. Five package ecosystems in two weeks. Each breach giving them the credentials to unlock the next one.

The payload was three stages.. harvest every SSH key, cloud token, Kubernetes secret, crypto wallet, and .env file on the machine.. deploy privileged containers across every node in the cluster.. install a persistent backdoor waiting for new instructions.

TeamPCP posted on Telegram after: “Many of your favourite security tools and open-source projects will be targeted in the months to come.. stay tuned.”

Every AI agent, copilot, and internal tool your company shipped this year runs on hundreds of packages exactly like this one… nobody chose to install LiteLLM on that developer’s machine. It came in as a dependency of a dependency of a plugin. One compromised maintainer account turned the entire trust chain into a credential harvesting operation across thousands of production environments in hours.

The companies deploying AI the fastest right now have the least visibility into what’s underneath it.

[Reda @redaelhail]

@Adam_USA1945 @lihoudii ها كيفاش كيربحو:
شي واحد عندو دين فالبنك 1000€
البنك كييأس وكيبيع هاد الدين لشي مستثمر غير بـ 100€
هاد المستثمر كيمشي عند الشخص اللي عليه الدين وكيحاول ياخذ منو 1000€
إلا قدر ياخذ حتى 400€ راه ربح، حيث هو شرا الدين غير بـ 100€
مرات كيمشيو للمحكمة ويقدرو يقطعو من الصالير

[Daniel Hnyk @hnykda]

LiteLLM HAS BEEN COMPROMISED, DO NOT UPDATE. We just discovered that LiteLLM pypi release 1.82.8. It has been compromised, it contains litellm_init.pth with base64 encoded instructions to send all the credentials it can find to remote server + self-replicate. link below

[News from Google @NewsFromGoogle]

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 more personalized Siri coming this year.

After careful evaluation, Apple determined that Google's Al technology provides the most capable foundation for Apple Foundation Models and is excited about the innovative new experiences it will unlock for Apple users. Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple's industry-leading privacy standards.