Last Week’s AI News #8

The AI world is spinning at full speed again, and the flow of news over the past few weeks has been simply massive. From companies increasing budgets for AI solutions to new startups and innovative technologies that have the potential to transform business, research, and creative processes, every development can have a real impact on small and medium-sized businesses (SMBs).

In this roundup, you’ll find the most important events from the past week that can help you understand AI trends, make strategic decisions, and see where technology evolution is heading. Whether you’re looking for ideas for your business or trying to understand market directions, this overview ensures you won’t miss the key moments.

Quick Highlights:

• AI is accelerating: companies are expanding budgets

• Human-centered AI: “Human&” seeks $1B investment to build next-generation AI

• Apple and Google Gemini: a new chapter in Siri evolution

• Sandbar launches Stream Ring – an AI ring converting whispers into notes

• Edison Scientific unveils “Kosmos” – autonomous AI for scientific research

• Alibaba-backed startup Moonshot AI introduces the Kimi K2 Thinking model

• Microsoft forms a new Superintelligence team

• OpenAI CFO retracts comments on federal guarantees

• Everything else that happened in AI last week

AI IS ACCELERATING: COMPANIES INCREASE BUDGETS

Wharton Business School released its annual corporate AI report, surveying around 800 US business leaders. The research shows that AI adoption is growing rapidly, budgets are expanding, and confidence in the technology is strengthening.

The most common AI use cases include data analysis, meeting summaries, reports, and marketing content creation. Popular tools include ChatGPT and Microsoft Copilot.

Why it matters for businesses:

• Companies that integrate AI quickly gain a competitive edge—they make faster decisions, save time, and create more value. While others are still weighing risks, early adopters already see clear returns on investment.

HUMAN-CENTERED AI: “HUMAN&” SEEKS $1B TO BUILD NEXT-GEN AI

Former xAI researcher Eric Zelikman is raising $1B in funding for the startup Human&, valued at $5B, to build human-centered AI using unique training methods. The team includes experts from OpenAI, Anthropic, Meta, DeepMind, and even Google’s seventh employee, Georges Harik.

Human& aims to create AI that better understands users and enhances their capabilities rather than replacing them. Zelikman is known for research on teaching language models to think step by step, which later helped develop OpenAI’s “o1” series.

Why it matters for businesses:

• This approach shifts the AI vision, from chasing “superintelligence” to building systems that help humans work smarter, faster, and more creatively. Such a perspective could fundamentally change how companies implement AI, from productivity to innovation. The high valuation before product launch also shows strong investor confidence in human-focused AI.

APPLE AND GOOGLE GEMINI: A NEW CHAPTER IN SIRI EVOLUTION

Bloomberg reports that Apple has partnered with Google to use the 1.2 trillion parameter Gemini model for an updated Siri version. The deal, worth around $1B per year, enables Gemini to handle summarization and planning tasks via Apple’s Private Cloud Compute system, ensuring data privacy.

Why it matters for businesses:

• Even giants like Apple choose collaboration over isolation. This demonstrates that speed and strategic alliances are key competitive advantages in the AI era.

SANDBAR LAUNCHES STREAM RING - AI RING CONVERTING WHISPERS INTO NOTES

Sandbar, a startup founded by former Meta designers, introduced the Stream Ring—an AI ring that captures quietly spoken thoughts and converts them into organized notes, while also functioning as a music controller.

The creators, Mina Fahmi and Kirak Hong, previously worked on neural interfaces at CTRL-Labs, acquired by Meta in 2019. Recording is activated by holding a touch zone, and whispers are detected by microphones and converted to text. The AI assistant responds in a synthesized voice, powered by ElevenLabs technology for natural conversation.

Why it matters for businesses:

• Another smart device demonstrates that voice AI technology is expanding into new formats—from rings to buttons to earbuds. Simple control can become an advantage, but competition among smart devices is intensifying rapidly.

EDISON SCIENTIFIC UNVEILS “KOSMOS”- AUTONOMOUS AI FOR SCIENTIFIC RESEARCH

Futurehouse launched its commercial division, Edison Scientific, and introduced “Kosmos”—an autonomous AI research system that can accomplish in one day what humans would take six months to do. Kosmos analyzes literature, processes data, and generates hypotheses, processing 1,500 articles and executing 42,000 lines of code per cycle, ensuring complete traceability. 79% of results have been validated, including unpublished discoveries across multiple scientific fields. Edison Scientific will commercialize the platform for the pharmaceutical sector, while Futurehouse continues fundamental non-profit projects.

Why it matters for businesses:

• Kosmos shows that AI can remove human time constraints in research, dramatically accelerate pace, and transform innovation speed across industries.

ALIBABA-BACKED MOONSHOT AI LAUNCHES KIMI K2 THINKING MODEL

Chinese startup Moonshot AI, backed by Alibaba, unveiled Kimi K2 Thinking—a open-source reasoning model that matches or surpasses models like GPT-5 and Claude 4.5 Sonnet at a fraction of the cost.

• Kimi outperformed GPT-5 and Sonnet 4.5 in several agent-based tasks, scoring 44.9% on Humanity’s Last Exam, setting a new record.

• The model improved significantly from its predecessor just four months ago, especially in coding, though it slightly trails the top leaders.

• Training costs were under $5M, with pricing far lower than current leading models.

Why it matters for businesses:

• Nvidia CEO Jensen Huang recently noted that China is only a few “nanoseconds” behind in AI. K2 Thinking shows that open-source and Chinese labs can compete with top models, offering a serious, affordable alternative to expensive closed solutions.

MICROSOFT FORMS NEW SUPERINTELLIGENCE TEAM

Microsoft AI CEO Mustafa Suleyman announced the MAI Superintelligence Team—a research division focused on building advanced systems to solve specific challenges in healthcare, energy, and other sectors, rather than general AGI.

• Suleyman emphasized a “humanistic superintelligence” vision, prioritizing AI that always works for people and humanity.

• The team will focus on narrow but high-impact societal challenges, including learning satellites, medical superintelligence, and clean energy solutions.

Why it matters for businesses:

• The new Superintelligence Team enables Microsoft to create AI solutions tackling real-world problems. Businesses can deploy advanced, human-beneficial technologies faster, strengthening competitive advantage and adapting to the rapidly evolving AI market.

OPENAI CFO RETRACTS COMMENTS ON FEDERAL GUARANTEES

OpenAI CFO Sarah Friar retracted comments suggesting the company seeks federal guarantees for infrastructure investments after heavy criticism. Sam Altman clarified that OpenAI opposes government bailout packages for private AI companies.

Initially, Friar told WSJ that OpenAI sought a federal “backstop” to fund AI investments, but later admitted she “misspoke.” White House AI chief David Sacks also rejected the idea, highlighting that other major AI companies could replace any failing competitor.

Altman stated that government guarantees for private AI projects are unnecessary, noting that OpenAI should face risk and accountability, while addressing criticism of the company being “too big to fail” and its massive expenses.

Why it matters for businesses:

• OpenAI’s spending and financial strategies have been closely scrutinized, and Friar’s comments sparked further debate. While unofficial, OpenAI’s investments, national infrastructure projects, and growing financial commitments raise important questions for the AI ecosystem and business landscape.

OTHER NOTABLE AI EVENTS LAST WEEK

• OpenAI hits 1M business customers: ChatGPT for Work grew 40% in two months, reaching 7M seats.

• Stability AI wins UK case against Getty Images: highlights challenges even large companies face in protecting creative work.

• Google integrates Gemini AI into Maps: enables conversational navigation and multi-step, building-aware directions.

• Snap partners with Perplexity: $400M investment to integrate AI into Snapchat by 2026, reaching nearly 1B users.

• Google retires Gemma model: due to factual inaccuracies; intended for developers and research.

• Cameo sues OpenAI over Sora feature: claims potential brand confusion with AI content.

• Udio sets 48-hour window for downloading AI-generated music following UMG partnership changes.

• OpenAI Sora allows purchasing extra generations: monetization pilot coming soon.

• AI music persona Xania Monet hits Billboard: first AI artist on radio charts after multimillion-dollar deal.

• Perplexity upgrades Comet AI assistant: improved web interaction, multi-tab support, and performance enhancements.

Last week demonstrated that AI is reshaping the business landscape - from productivity solutions to next-gen research and creative content. Every new development can directly impact companies striving to stay competitive.

Want to be the first to know the latest AI trends and how to apply them to your business? Subscribe to our weekly AI newsletter and get the most important insights delivered straight to your inbox!

Don't miss news from us

Or join AI Mastery Program for business leaders >