💻 GitHub launches AI-powered app builder Spark 💻
Copilot Pro+ users can now build full-stack apps in minutes

GitHub just launched Spark in public preview, a revolutionary tool that lets developers build and deploy full-stack AI apps using nothing but natural language descriptions. First announced last year, Spark uses Anthropic's Claude Sonnet 4 model to field users' requests and automatically generate complete applications with frontend, backend, and hosting included. The platform is available exclusively to Copilot Pro+ subscribers and represents a major leap in the "vibe coding" trend that's reshaping software development.

Microsoft CEO Satya Nadella announced the launch on Twitter, describing Spark as "a new tool in Copilot that turns your ideas into full-stack apps, entirely in natural language." The platform requires no setup or configuration, allowing users to go from concept to deployed application in minutes. This positions GitHub to compete directly with emerging no-code platforms like Lovable, which recently crossed $100M ARR in just eight months - faster than any other software company in history.

The launch comes as Microsoft doubles down on AI-powered development tools, with 150 million+ programmers already using GitHub. Spark's "Dream it. See it. Ship it." tagline reflects the company's ambition to democratize app development and capture the growing market of citizen developers who want to build software without traditional coding skills.

🎯 Microsoft hires 24 Google DeepMind employees🎯
24 AI experts including Gemini VP join Microsoft's AI division

Microsoft has successfully poached approximately 24 employees from Google's DeepMind AI lab in recent months, marking the latest escalation in the AI talent war between tech giants. The most notable hire is Amar Subramanya, who spent 16 years at Google as VP of engineering developing Gemini before joining Microsoft in June as Corporate VP in its AI division. Adam Sadovsky also made the jump after 18 years at Google as a software engineer and senior director at DeepMind.

The aggressive hiring spree follows Meta's strategy of targeting top researchers with up to $100 million signing bonuses and one-year compensation packages. Microsoft's talent acquisition is particularly striking given that DeepMind uses aggressive noncompete clauses and up to a year-long PTO to prevent staffers from joining rival firms. The company laid off more than 9,000 people from its workforce to free up spending money for AI infrastructure and talent poaching.

This hiring blitz appears driven by Microsoft's Copilot struggles. High-ranking executives have described Copilot as "gimmicky" and admit it "isn't as good as ChatGPT." The company outsources help from third-party vendors to get Copilot to deliver value, while corporations remain more inclined toward ChatGPT because "OpenAI has done a tremendous job making their product fun to use." The DeepMind talent acquisition represents Microsoft's attempt to close this capability gap.

🇺🇸 Trump Unveils AI Action Plan Targeting Deregulation 🇺🇸
Anti-"woke AI" executive orders aim to cement US dominance

President Trump unveiled his comprehensive AI Action Plan on July 23, a 28-page strategy called "Winning the Race: America's AI Action Plan" that centers on three pillars: accelerating AI innovation, building AI infrastructure, and leading international AI diplomacy. The plan includes over 90 federal actions focused on deregulation and removing what Trump calls "red tape" that could slow AI development in the race against China.

The administration signed three AI-focused executive orders, including one that bans federal agencies from contracting with tech companies that operate AI chatbots displaying "partisan bias" - defined as diversity, equity, and inclusion considerations. Trump declared there will be "no woke AI in Washington" and vowed to turn the US into an "AI export powerhouse" by loosening environmental rules and vastly expanding AI exports to allies.

The plan fast-tracks federal permitting and streamlines reviews to "do everything possible to expedite construction of all major AI infrastructure projects." It also threatens to limit federal funding for states that pass AI laws deemed "burdensome" to developing the technology. The approach marks a sharp departure from the Biden administration's focus on AI safety and regulation, instead prioritizing speed and competitiveness in the global AI race.

INTERESTING TO KNOW

🚨 $1 Billion Worth of Nvidia AI Chips Smuggled to China Despite Export Controls 🚨

At least $1 billion worth of Nvidia's advanced AI processors were smuggled to China in the three months after Donald Trump tightened chip export controls, according to a Financial Times investigation. The report reveals that Nvidia's B200 has become the most sought-after processor on China's black market, with these chips being critical for training AI systems like those used by OpenAI and Google. Despite national security concerns and strict export bans, sophisticated smuggling networks continue to operate, with one source describing the underground market as being "like a seafood market" in terms of availability. The smuggling operation highlights the challenges of enforcing tech export controls and demonstrates China's determination to access cutting-edge AI hardware despite US restrictions.

🗣️ Boson AI releases open-source voice cloning model🗣️

California-based Boson AI just released Higgs Audio V2, a surprisingly powerful open-source text-to-speech model that's capable of voice cloning and generating multi-speaker conversations with frontier-level performance. The model was pre-trained on over 10 million hours of audio data and demonstrates remarkable capabilities in one video demo where a Korean speaker and an English speaker have a seamless conversation by using V2 to translate their voices in near real-time. The open-source release positions Boson AI as a serious competitor to established players like ElevenLabs and OpenAI's voice models, while offering developers free access to state-of-the-art voice synthesis technology. The model supports zero-shot voice cloning and can run on edge devices, making it particularly attractive for developers building voice-enabled applications.