The most important news pieces from the AI world are here for you, whether you’re an AI enthusiast or just wish to keep an eye on what’s going on.
Focus Topic: Apple’s AI delay shows the cost of overpromising
Apple’s delayed Siri features are a useful reminder that AI demos and real AI products are not the same thing.
According to this week’s news, Apple agreed to pay $250 million in a class-action settlement related to the delayed personalized Siri features first introduced at WWDC 2024. Eligible iPhone buyers in the US may receive around $25 per device, with the amount potentially rising up to $95. The lawsuit argued that Apple promoted AI features that were not actually available at the time they were announced.
This matters because the AI market has become very good at creating expectations. Every company wants to show that it has an AI assistant, an AI agent or an AI-powered operating system. But once these features move from keynote slides into people’s phones, the bar becomes much higher. The product must understand context, access personal data safely, work across apps and avoid making mistakes in sensitive situations.
Apple’s situation also shows why “personal AI” is one of the hardest AI product categories. A general chatbot can say “I’m sorry, I can’t help with that”. But a personal assistant that reads your messages, edits your photos, books your calendar or changes device settings has less room for failure. It needs trust, reliability and privacy at the same time.
The bigger lesson is simple: in AI, launching late may hurt the brand – but launching too early can be even more expensive. As Apple prepares new Siri features for iOS 27, the real question is whether the next generation of AI assistants can finally match the promises made on stage.
LLMs & AI Models
- OpenAI introduced GPT-Realtime models that bring GPT-5-level reasoning, longer context and live translation to voice agents.
- A Harvard study found OpenAI’s o1-preview outperformed ER doctors in controlled diagnosis tests on real emergency cases.
- OpenAI launched GPT-5.5 Instant as ChatGPT’s default model, with fewer hallucinations and stronger personalization.
- Anthropic introduced 10 ready-made AI agents for finance, covering KYC checks, earnings analysis and pitch book creation.
New Tools
- OpenAI is accelerating its AI-agent phone, aiming for mass production in 2027 and up to 30 million devices by 2028.
- Apple plans to open iOS 27 to rival AI models, letting users choose which service powers Siri and other AI features.
- Spotify is testing a tool that lets AI agents create and upload personal podcasts for private listening.
- Google is testing Remy, a proactive 24/7 Gemini agent that can use data across Gmail, Calendar, Photos and other apps.
- Anthropic signed a compute deal with SpaceX to use more than 220,000 Nvidia GPUs for expanding Claude Code capacity.
👉 Explore these tools: Spotify AI podcasts | Claude Code
Other Quick Picks
- Jack Clark predicts a self-improving AI system could arrive before 2029, with over 60% probability.
- Google quietly shut down Project Mariner, moving parts of the browser-agent tech into Gemini and Chrome.
- Google is turning Fitbit into Google Health, adding a Gemini-based AI health coach and new Fitbit Air device.
- Peter Thiel led a $140M round in Panthalassa, which builds floating AI data centers powered by ocean waves.
- Mira Murati testified that Sam Altman was not always candid during OpenAI’s 2023 board crisis.
- The White House is considering stricter pre-release reviews for high-risk AI models after Anthropic Mythos.
- Coinbase cut 14% of staff, saying AI lets smaller engineering teams work faster.
- Greg Brockman disclosed an OpenAI stake valued near $30B during testimony in Musk’s lawsuit.
- DeepMind UK employees voted to unionize after Google’s classified Pentagon AI contract.
- Meta will use AI to detect underage users from photos, videos and profile context in the EU and Brazil.
- Anthropic, Goldman Sachs and Blackstone formed a $1.5B venture to bring AI into portfolio companies.
- Microsoft, Google and xAI will let the US government review new AI models before release.
🇪🇪 AI News from Estonia
- A London startup with Estonian unicorn roots is opening a Tallinn development center where employees work alongside AI agents.
- Estonian companies face legal and business risks if AI is used without proper governance, transparency and documentation.
- Oskar Gross and Ago Ambur are building Glazer, an AI tool that turns weeks of investigative work into minutes.
- Ott Velsberg left his role as Estonia’s chief data officer and moved to a new international role in London.
- Elisa’s group lead Topi Manner says AI has reduced the company’s need for outsourced coding work from India.
- An Estonian startup OÜ HotelBuddy Technology partnered with a hotel group operating in 110 countries with an AI-based hospitality solution.
- Villmann shared how a campaign for the Consumer Protection and Technical Regulatory Authority was made fully with AI.
- Tilde is building AI language technology for Baltic and smaller languages to protect them in the AI era.
- The European Commission opened consultation on AI Act transparency rules, giving Estonian organizations a chance to respond.
🎙️ AIPowerment Podcast dropped a new month-end episode of AI news covering GPT-5.5, Anthropic leaks, and AI’s impact on jobs and society.
🎧 Listen to Inside Finance Podcast on Spotify, Apple Podcasts, and YouTube.
Want to stay in the loop?
Straight to your inbox: practical AI updates, finance use cases, tools to try, and upcoming trainings.
I cut the noise and send only what’s actually useful.
💼 Exploring how to make AI work in finance? Let’s talk practical use cases – connect with Gerlyn Tiigemäe for expert guidance.
📚 If you would like to participate in one of my trainings or listen to speaking engagements, here are the upcoming ones:
- Training “AI võimalused finantsvaldkonnas” @ Äripäeva Akadeemia, next trainings for beginners @ 14.05 – information; for intermediate level @ 05.06 – information
- Training “AI võimalused projektijuhtidele” @ Äripäeva Akadeemia, next trainings @ 15.06 – information
