Welcome to the first of my regular columns, which keep you up to speed with the latest artificial intelligence news, focusing on informing novice and laypeople. For the past 84 weeks, I’ve manually organized around 500 AI headlines per week into 45 categories.
That’s over 28,000 organized links.
This has given me an odd ‘spidey sense’ of what’s happening, and I’m excited to share it with you.
We’re in a chapter of technology where the sum of the parts will become greater than the whole. It’s like the early days of the Internet. While it’s tempting to seek surgical examples of “AI for journalism,” it’s equally important to understand the big picture to see and prepare for opportunities and challenges as they emerge.
To balance this, the column will begin with the broad top stories and then shift into publishing-related stories.
Audacious AI headlines to supplement our morning coffee
Anthropic has doubled down on its prediction that we will achieve Nobel Prize-level artificial intelligence across domains at scale within three years.
“Worth noting that Anthropic is not backing down on the timeframe for having AIs with ‘Intellectual capabilities matching or exceeding that of Nobel Prize winners across most disciplines—including biology, computer science, mathematics, and engineering.’ …in the next three.” / X
https://x.com/emollick/status/1917698375115321714
Microsoft predicts that 2025 will be the year when “frontier firms” emerge and AI employees begin to take the lead on operational tasks.
2025: The year the frontier firm is born
https://www.microsoft.com/en-us/worklab/work-trend-index/2025-the-year-the-frontier-firm-is-born
Anthropic predicts fully automated AI employees within the next 12 months.
Exclusive: fully AI employees are a year away, Anthropic warns
https://www.axios.com/2025/04/22/ai-anthropic-virtual-employees-security
Demis Hassabis, CEO of Google DeepMind, forecasts that AI will naturally evolve self-awareness within the next 5 to 10 years.
Demis Hassabis says today’s AI lacks consciousness, but self-awareness could emerge “implicitly.” Models may soon need to understand “you, self and other” — the early elements of awareness. They lack imagination now, but in five to 10 years, Hassabis predicts, they’ll solve and pose scientific conjectures.
Anthropic is urging immediate action on chip exports due to concerns of international conflict stemming from potential 2026 AI breakthroughs.
Anthropic’s AI export controls framework response
https://www.anthropic.com/news/securing-america-s-compute-advantage-anthropic-s-position-on-the-diffusion-rule
Morgan Stanley forecasts 1 billion humanoid robots by 2050.
“Morgan Stanley forecasts a $4.7 trillion global annual humanoid robot revenue by 2050, with roughly 1 billion units cumulatively deployed. That sounds incredibly bearish – something would have to go seriously wrong with the world to derail momentum to that extent.”
Grounded AI headlines about things happening right now
AI agents
Yelp began testing AI-powered voice agents to help restaurants and businesses answer and manage incoming phone calls.
Yelp debuts AI voice agents for restaurants and service providers | TechCrunch
https://techcrunch.com/2025/04/29/yelp-is-adding-ai-powered-voice-agents-for-restaurants-and-services/
Early preview of Yelp’s upcoming AI-powered call answering services – YouTube https://www.youtube.com/watch?v=OboONaC4WkA
ElevenLabs, the audio AI company, has developed a method to allow voice agents to transfer calls to other voice agents. That’s a spectacular development and it implies that different specialists could take on different tasks and transfer those tasks like humans.
Agent transfer: Seamlessly transfer the user between Conversational AI agents based on defined conditions.
https://elevenlabs.io/docs/conversational-ai/customization/tools/system-tools/agent-transfer
OpenAI launched a GPT-powered WhatsApp phone number to let people text or call to get answers in real time through their phones. I think this is an underreported story because it is a direct harbinger of a shift towards talking to AI agents as if we would talk to people. If you’ve ever seen the viral chart of how often people go to church or interact with their friends in person, I think we’re going to see a new line on that chart where AI agents replace a significant portion of our interaction with others.
Search in WhatsApp: You can now send a WhatsApp message to 1-800-ChatGPT (+1-800-242-8478) to get up-to-date answers and live sports scores. Accessible everywhere ChatGPT is available. https://x.com/OpenAI/status/1916947244852646202
Meta launched a standalone app that has been billed as providing companionship. I’m not a fan.
Introducing the Meta AI App: A new way to access your AI assistant
The Harvard Business Review released a study that shows that people are using AI predominantly for emotional and everyday support. Therapy and companionship is the number-one personal use case of artificial intelligence.
How people are really using Gen AI in 2025
https://hbr.org/2025/04/how-people-are-really-using-gen-ai-in-2025
Coding assistance
In separate statements, Microsoft CEO Satya Nadella and Google CEO Sundar Pichai revealed that AI now generates over 30% of all code at their companies.
Satya Nadella says as much as 30% of Microsoft code is written by AI. https://www.cnbc.com/2025/04/29/satya-nadella-says-as-much-as-30percent-of-microsoft-code-is-written-by-ai.html
Sundar Pichai said, “I think the last time I had said the number was like 25% of code that’s checked in involves people accepting AI suggested solutions. That number is well over 30% now.”
https://seekingalpha.com/article/4777993-alphabet-inc-goog-q1-2025-earnings-call-transcript
Cursor, an AI coding assistant, is now generating 1 billion lines of accepted code every day.
“Cursor writes almost 1 billion lines of accepted code a day. To put it in perspective, the entire world produces just a few billion lines a day.”
https://x.com/amanrsanger/status/1916968123535880684
Uber is using AI agents to deploy code across the company at an enterprise level.
To tackle large-scale code migrations, Uber’s Developer Platform team uses LangGraph to build a network of agents and automate unit test generation. Uber has established a dedicated Developer Platform AI team to more deeply integrate agentic systems into their engineering workflows. Using LangGraph, the team is building reusable components that can be adopted across the broader organization—starting with automating unit test generation as a key use case.
OpenAI came out with a 34-page guide to explain how to make autonomous agents.
A practical guide to building agents
https://cdn.openai.com/business-guides-and-resources/a-practical-guide-to-building-agents.pdf
Shopping and Payments
OpenAI rolled out enhanced ChatGPT shopping. I’m not sure if I wish I was back in retail or not based on this one. I think it’s actually going to be quite fun for retailers to figure out a new world, and I think it’s going to be especially valuable for consumers who won’t have to deal with incessant marketing garbage. Just as ratings and reviews leveled the playing field for transparency and the voice of the customer, I think artificial intelligence is going to get rid of the sea of nonsense on the Internet that exists only to push things on us. If anything, we’ll see a resurgence back to the world of XML and product feeds.
Improved shopping results from ChatGPT search
https://help.openai.com/en/articles/11146633-improved-shopping-results-from-chatgpt-search
“We’re excited to announce we’ve launched several improvements to ChatGPT search, and today we’re starting to roll out a better shopping experience. Search has become one of our most popular & fastest growing features, with over 1 billion web searches just in the past week 🧵” / X
https://x.com/OpenAI/status/1916947241086095434
OpenAI upgrades ChatGPT search with shopping features | TechCrunch https://techcrunch.com/2025/04/28/openai-upgrades-chatgpt-search-with-shopping-features/
“Shopping in ChatGPT (including direct links to buy) rolling out starting today:” / X https://x.com/gdb/status/1917009041035038837
Mastercard and Visa separately introduced AI agent integration into their payment systems to automate AI purchasing.
Mastercard unveils Agent Pay, pioneering agentic payments technology to power commerce in the age of AI | Mastercard Newsroom https://www.mastercard.com/news/press/2025/april/mastercard-unveils-agent-pay-pioneering-agentic-payments-technology-to-power-commerce-in-the-age-of-ai/
Find and buy with AI: Visa unveils new era of commerce https://www.businesswire.com/news/home/20250430580204/en/Find-and-Buy-with-AI-Visa-Unveils-New-Era-of-Commerce
Sounds Like Science Fiction
Waymo has reached 250,000 weekly paid robo taxi rides across the United States in four cities.
“We’re now providing more than 250,000 fully autonomous paid rides each and every week. The robotaxi future is here, and it’s powered by our generalizable Waymo Driver” https://x.com/Waymo/status/1915507165378257100
In ethics and security news, a new benchmark has shown that OpenAI’s O3 model can outperform 94 percent of expert virologists. That’s relaxing.
Exclusive: AI bests virus experts, raising biohazard fears | TIME
https://time.com/7279010/ai-virus-lab-biohazard-study
Recent AI news with real publisher impact
The end of the internet user interface as we know it
Google announced that its AI-enhanced search now reaches 1.5 billion users per month.
Google’s AI overviews reach 1.5 billion monthly users https://www.searchenginejournal.com/googles-ai-overviews-reach-1-5-billion-monthly-users/545333/
A study showed that AI reviews reduce click-through rates by 34%.
AI overviews reduce clicks by 34.5%
https://ahrefs.com/blog/ai-overviews-reduce-clicks/
Andrej Karpathy, one of the most generous, smartest, kindest, and broad-visioned voices in AI (and I don’t say that lightly), co-founder of OpenAI and former head of AI at Tesla, has said that our primary audience as product developers, publishers and e-commerce website developers is no longer human. Instead, our primary audience is going to be web crawlers and agents that are looking on behalf of people who have asked them questions. That’s a completely different usability assignment than traditional user experience. I agree completely with this statement. I’m actually quite excited about this.
“PSA It’s a new era of ergonomics. The primary audience of your thing (product, service, library, …) is now an LLM, not a human. LLMs don’t like to navigate, they like to scrape. LLMs don’t like to see, they like to read. LLMs don’t like to click, they like to curl. Etc etc.” / X
https://x.com/karpathy/status/1914494203696177444
“Tired: elaborate docs pages for your product/service/library with fancy color palettes, branding, animations, transitions, dark mode, … Wired: one single docs .md file and a “copy to clipboard” button.” / X https://x.com/karpathy/status/1914488029873627597
OpenAI has expressed interest in purchasing the Google Chrome browser, if given the chance.
OpenAI would buy Google’s Chrome, exec testifies at trial | Reuters https://www.reuters.com/sustainability/boards-policy-regulation/google-contemplated-exclusive-gemini-ai-deals-with-android-makers-2025-04-22/
Microsoft has added the ability to “see” to its web browser. It also launched computer use and web browsing within the operating system. This is a pretty big first step towards the end of the Internet as we know it.
Microsoft Edge on X: “We just leveled up! Introducing @Copilot Vision in Edge. It can summarize aloud and even help you practice for a job interview. It’s free to use – open Copilot in Edge, click on the mic icon, and it’ll see what you see. https://t.co/wEnCN9vfmA” / X
Anthropic launched a research feature. This means that Google, OpenAI, Perplexity, Claude, and DeepSeek all now have robust research capabilities. This is also a harbinger for the end of the Internet as we know it.
Claude takes research to new places:
https://www.anthropic.com/news/research
AI ethics and legal updates
The Washington Post has partnered with OpenAI to include news within ChatGPT chats.
The Washington Post partners with OpenAI on search content – The Washington Post
https://www.washingtonpost.com/pr/2025/04/22/washington-post-partners-with-openai-search-content
The Oscars officially OK’d the use of artificial intelligence. They don’t even require disclosure. Talk about a change of heart in less than 12 months.
Oscars OK the use of AI, with caveats – The New York Times https://www.nytimes.com/2025/04/21/business/oscars-rules-ai.html
Adobe has now incorporated Google and OpenAI’s models into its Firefly product. That’s another pretty big shift in ethics.
Adobe adds AI models from OpenAI, Google to its Firefly app | Reuters https://www.reuters.com/business/adobe-adds-ai-models-openai-google-its-firefly-app-2025-04-24/
Things to share with your IT department
Nvidia came out with four models of their own that could help publishers:
- Describe Anything 3B is able to create highly detailed descriptions of images and videos as well as caption them. It’s especially good at diving into specific elements of an image or a video to give extra detail.
- Eagle 2.5 is a small model that has been created for understanding extremely long videos and complex imagery. It’s also really good at understanding charts and graphs.
- Nemotron-UltraLong-8B was created with a huge memory and supports up to 4 million tokens. That means an entire enterprise code base could be loaded into it and analyzed.
- Parakeet: a new standard for automatic speech recognition (ASR) with an industry-best 6.05% Word Error Rate on the HuggingFace Open-ASR-Leaderboard.
It’s important to remember that Nvidia is making models as well as chips – and most are open source.
Ethan Holland is on the Local Media Association’s board of directors. He is the vice president of Draper Digital Media.