Hello Engineering Leaders and AI Enthusiasts!
This newsletter brings you the latest AI updates in a crisp manner! Dive in for a quick recap of everything important that happened around AI in the past two weeks.
And a huge shoutout to our amazing readers. We appreciate youš
In todayās edition:
ā ļø Appleās Siri upgrades delayed amid internal challenges
š China has another DeepSeek moment with Manus
š§Ŗ OpenAI is pushing the next wave of AI agents
š± Deutsche Telekomās “AI Phone” in Partnership with Perplexity
š¤ Alibaba’s new model outperforms DeepSeek-R1
āļø Amazonās Nova: A new rival to OpenAI and Anthropic
šļø OpenAI pushes for federal AI regulations
š§ Knowledge Nugget: AI for PMs: Disruption or Partnership? by
Letās go!
Appleās Siri upgrades delayed amid internal challenges
Apple has officially delayed its Siri and AI enhancements, initially promised last June, with no clear timeline for release. In a Siri team meeting led by senior director Robby Walker, he acknowledged the āuglyā situation, employee frustration, and uncertainty about whether the missing features will make it into iOS 19.
The delay stems from quality issues as well as competing internal priorities taking precedence. However, Walker reassured that Siri improvements will ship āas soon as they are ready,ā emphasizing Appleās commitment to quality over rushing features to market.
Why does it matter?
Apple has been unable to escape headlines about its slow progress with everything having to do with Siri and artificial intelligence. This casts further doubts on its ability to compete with OpenAI, Google, and others. Could it also impact Apple usersā trust and market perception?
China has another DeepSeek moment with Manus
A Chinese startup has introduced Manus, claiming it to be the worldās first fully autonomous AI agent. Manus can independently handle real-world tasks and has achieved state-of-the-art (SOTA) performance on agentic benchmarks.
A demo by Manus AI chief scientist Ji showcased Manus completing three distinct tasks: sorting resumes, analyzing stock correlations, and evaluating New York real estate by assessing nearby schools and affordability. However, Manus remains in invitation-only private testing for now.
Following its recent viral success, the company has announced a strategic partnership with Alibabaās Qwen team to develop a Chinese version of its autonomous agent platform.
Why does it matter?
Manus is fueling concerns that China is rapidly advancing in AI, drawing comparisons to DeepSeek and raising fears of a growing AI rivalry. While some praise its capabilities, others question its limitations and potential risks to privacy and security. Autonomous AI agents were already a major industry trend, and Manusā debut has intensified the discussion.
OpenAI is pushing the next wave of AI agents
OpenAI introduced new tools that allow businesses to create their own custom AI agents capable of handling tasks like web browsing and file management, marking a major step toward autonomous enterprise AI.
The company may also be planning to launch a tiered suite of specialized AI agents, priced between $2,000 and $20,000 per month, catering to business professionals, advanced developers, and Ph.D.-level researchers. SoftBank has reportedly committed $3 billion to these AI agents for 2025 alone, with OpenAI expecting them to contribute up to 25% of its long-term revenue.
Why does it matter?
OpenAIās expansion into AI agents marks a major step toward autonomous enterprise AI. It’s pricing could set new benchmarks for AI agent costs while also shedding some light on how much businesses are willing to invest in AI-powered automation.
Deutsche Telekomās “AI Phone” in Partnership with Perplexity
T-Mobile’s parent company, Deutsche Telekom, announced development of an “AI Phone”. Running on the custom Magenta AI OS, the device will integrate Perplexity Assistant and other services like Google Cloud AI, ElevenLabs, and Picsart, supporting voice, text, and camera inputs for various web tasks. It is expected to launch in Europe by late 2025 for under $1,000, with a global rollout in 2026.
Why does it matter?
This marks a major step toward AI-native smartphones, where AI isn’t just an app but a core part of the deviceās functionality. It could set a new trend for AI-driven devices, influencing the future of the smartphone industry.
Alibaba’s new model outperforms DeepSeek-R1
Alibaba launched QwQ-32B, an open-source reasoning model that matches the performance of DeepSeekās R1 (with 671 billion parameters) while requiring significantly less computing power. This explores how reinforcement learning enhances foundation models pretrained on vast knowledge. The model also has agent-like capabilities, allowing it to think critically, use tools, and refine its reasoning based on environmental feedback.
Why does it matter?
This release underscores the rapid evolution of reasoning models from cutting-edge breakthroughs to widely accessible AI tools. The modelās efficiency also signals a shift toward making high-performance AI more cost-effective and scalable.
Amazonās Nova: A new rival to OpenAI and Anthropic
Amazon is developing Nova, an advanced reasoning AI model set to launch by June, aiming to rival OpenAIās o1 and Anthropicās Claude 3.7 Sonnet. Nova features a āhybrid reasoningā approach, balancing speed with deep analytical thinking, and focuses on software development and math benchmarks.
Amazon is positioning Nova as a cost-effective yet high-performing alternative, marking its biggest push into AI competition despite its $8 billion investment in Anthropic.
Why does it matter?
If successful, Nova could help reshape the competitive landscape, offering businesses an alternative for software development, math, and analytical tasks while strengthening AWSās AI ecosystem.
OpenAI pushes for federal AI regulations
OpenAI has submitted a 15-page proposal to the U.S. government advocating for federal-level AI regulations to replace inconsistent state laws, arguing that regulatory clarity is essential for U.S. AI competitiveness against China.
The company is also advocating for fair use of copyrighted materials in AI training, suggesting that AI models need broad data access to drive innovation. Additionally, it is calling for a faster approval process for AI tools and proposing restrictions on Chinese AI models like DeepSeek.
Why does it matter?
OpenAIās push for stronger federal AI policies could shape the future of global AI leadership, balancing innovation, security, and economic growth while setting a precedent for AI governance worldwide.
Enjoying the latest AI updates?
Refer your pals to subscribe to our newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the āShareā button on any post, you’ll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: AI for PMs: Disruption or Partnership?
In this insightful post, discusses how AI is reshaping product management, but not by eliminating the roleārather by shifting its focus. While AI can automate execution-heavy tasks like roadmap updates, PRDs, and backlog prioritization, human judgment remains critical for strategy, stakeholder alignment, and decision-making.
As AI expands PMsā capacity, companies will need fewer entry-level PMs, making AI adoption essential for career growth. The future of PM work is not AI vs. humansāitās AI-empowered PMs driving bigger initiatives and managing greater complexity.
Why does it matter?
The post is a reminder that it’s not longer just about AI replacing jobsāitās about people who use AI outperforming those who donāt. AI is becoming a competitive advantage across industries, and those who fail to adapt risk falling behind.
What Else Is Happeningā
š°DeepSeek claims its AI models achieve a 545% profit margin on inference costs, sharply contrasting with U.S. competitors that are running at a loss.
šØāš»Anthropic is āreservingā Claude 4 models for āsubstantial leapsā, predicting AI will surpass top coders by 2026.
šSamsung unveiled new $300 Galaxy A series phones with AI features like Circle to Search and AI photo editing, aiming to compete with Apple’s $599 iPhone 16e.
šOpenAI rolled out GPT-4.5-Preview model to all Plus users and hinted at integrating its AI video tool, Sora, into ChatGPT.
š”Google is launching new āAI Modeā experimental feature directly in Search, allowing users to ask complex, multi-part questions and follow-ups to dig deeper on topics.
šMistral AI launched Mistral OCR, a document analysis API that processes text, tables, media, and equations, with exceptional accuracy and speed.
š¤Microsoft is developing MAI, a new AI model family, and in-house reasoning models to reduce reliance on OpenAI.
š§ Tencent launched Hunyuan-TurboS, an advanced AI model mimicking human cognitive processesāfast thinking and slow thinkingāfor better efficiency and reasoning.
š„ļøMeta is testing its first in-house AI training chip to boost AI capabilities and reduce dependency on external suppliers.
ā”Google introduced Gemma 3, a lightweight AI model rivaling larger ones, and launched Gemini Robotics for smarter robots. It will also replace Google Assistant with Gemini later this year.
šCohere unveiled Command A, an enterprise AI model with top-tier performance on just two GPUs and strong multilingual support.
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you next week! š
Read MoreĀ in Ā The AI EdgeĀ
