Nahda

June 27, 2026

OpenAI unveils its first AI chip as rivals push Claude and Gemini deeper into daily work

A promotional image from OpenAI's announcement of Jalapeno, its first custom-designed AI chip.

AI moved closer to your desk this week

This week the AI story moved out of the research lab and into the real world. OpenAI designed its own chip to make its tools cheaper, Google's DeepMind put $75 million behind AI tools made for filmmakers, and Anthropic dropped Claude straight into the chat app where work already happens, even as it picked a public fight over who is copying whom. Underneath the noise, one quiet thread runs all the way back a thousand years to Cairo: progress comes from testing things for yourself.

OpenAI unveils Jalapeno, its first custom AI chip built with Broadcom

An image from OpenAI's announcement of its first custom AI chip, Jalapeno.

OpenAI showed off Jalapeno, the first computer chip it designed itself, built with Broadcom and aimed at running its models more cheaply. Today OpenAI rents most of its computing power from Nvidia, so making its own chip is a bet that it can cut the enormous cost of answering billions of ChatGPT requests. For everyone else, cheaper computing usually means faster, lower-priced AI tools over time. The first chips are expected to come online by the end of 2026.

Source: TechCrunch

Google's DeepMind invests $75 million in A24 to build AI filmmaking tools

A film production setting, illustrating the A24 and Google DeepMind partnership to build AI filmmaking tools

Google's AI lab DeepMind is putting $75 million into a multi-year partnership with A24, the studio behind films like Everything Everywhere All at Once, to build AI tools made specifically for filmmakers. Early prototypes include storyboarding tools, and DeepMind's video generator Veo is expected to play a big role. The deal matters because it pushes AI in film past generic video generators toward tools shaped by people who actually make movies, with anything built also flowing back into Google's products. The partnership is non-exclusive, so A24 can still work with other AI companies too.

Source: The Hollywood Reporter

Anthropic puts Claude inside Slack as a teammate you can tag in

A screenshot of Anthropic's Claude Tag feature working inside a Slack channel.

Anthropic put its Claude assistant directly inside Slack, the chat app many teams already use for work. Anyone in a channel can now type @Claude to hand it a task, and it remembers the channel's context so you do not have to re-explain things every time. Anthropic says Claude already writes about 65 percent of its own product team's code this way. It is an early look at what it means to treat an AI less like a search box and more like a coworker.

Source: TechCrunch

Anthropic publicly accuses Alibaba of secretly copying Claude

The Alibaba logo displayed at a company event.

Anthropic publicly accused China's Alibaba of secretly copying its Claude models, in a letter it called the largest such case it has seen. The claim centers on a practice called distillation, where one company quietly uses another's AI to train a cheaper imitation of it. Alibaba had not publicly addressed the specific accusations. The dispute is a sign of how valuable these models have become and how hard they are to protect.

Source: CNBC

Arab startups keep raising as Foodics buys an AI firm and Morocco's Agenz lands $5 million

A business scene illustrating the article on Middle East and North Africa startup funding.

Money kept flowing to startups across the Arab world this week. Saudi Arabia's Foodics, which makes software for restaurants, fully acquired the Greek AI company Norma to build smarter tools into its product, while Morocco's property platform Agenz raised $5 million and the UAE's Revora raised $2 million. The deals stretch from the Gulf to North Africa, a reminder that the region's tech scene is broad and still growing. For founders here, it is a sign that regional investors are active and that local problems are worth building for.

Source: Arab News

Ibn al-Haytham settled an argument about sight with a dark room and a pinhole

Around a thousand years ago in Cairo, the scholar Ibn al-Haytham set out to settle a long-running argument: do we see because the eye sends out rays, as the ancient Greeks taught, or because light enters the eye? Instead of trusting the old authorities, he darkened a room, let light pass through a tiny hole, and studied how images formed on the far wall, showing that sight comes from light coming in. He wrote that a true seeker should doubt the books of the ancients and put their claims to the test. This is the heart of the scientific method: do not accept a claim because someone important said it, settle it with an experiment you can repeat. Try it: this week, take one thing you are sure is true at work and design one small, cheap test that could prove you wrong.

Source: Wikipedia

We are looking for Syrian projects to feature

In every issue we like to highlight something being built in Syria, or by Syrians anywhere: a product, a startup, a tool, or an idea in its early days. If you are building something, or know someone who is, tell us about it. Tell us what you are making, who it is for, and where you are today. Write to us, and your project might be the story in the next issue.

The board: jobs and an invitation to come home

This is a space to connect people with opportunities. If your team is hiring in tech, AI, or entrepreneurship, or you are a Syrian thinking about coming home and looking for a place to build, this board is for you. Send us the role or what you are looking for in one clear sentence, and we will help get it to the people who need it. Let us build this bridge together.

The newest tools rest on the oldest habit

Every chip and model in this issue is, at bottom, a faster way to ask a question and check the answer. That habit is a thousand years old. Keep asking.