The US AI Action Plan, Explained— from theneurondaily.com by Grant Harvey Sam’s 3 AI nightmares, Google hits 2B users, and Trump bans “woke” AI…
Meanwhile, at the Fed’s banking conference on Wednesday, Altman revealed his three nightmare AI scenarios. The first two were predictable: bad actors getting superintelligence first, and the classic “I’m afraid I can’t do that, Dave” situation.
But the third? AI accidentally steering us off course while we just…go along with it.
His example hit home: young people who can’t make decisions without ChatGPT (according to Sam, this is literally a thing). See, even when AI gives great advice, collectively handing over all decision-making feels “bad and dangerous” (even to Sam, who MADE this thing).
So yeah, Sam’s not really worried about the AI rebelling. He’s worried about AI becoming so good that we stop thinking for ourselves—and that might be scarier.
Also from The Neuron re: the environmental impacts of producing/offering AI:
In Episode 5 of The Neuron Podcast, Corey Noles and Grant Harvey tackle the education crisis head-on. We explore the viral UCLA “CheatGPT” controversy, MIT’s concerning brain study, and innovative solutions like Alpha School’s 2-hour learning model. Plus, we break down OpenAI’s new $10M teacher training initiative and share practical tips for using AI to enhance learning rather than shortcut it. Whether you’re a student, teacher, or parent, you’ll leave with actionable insights on the future of education.
“Our data is showing that 70 percent of the skills in the average job will have changed by 2030,” said Aneesh Raman, LinkedIn’s chief economic opportunity officer. According to the World Economic Forum’s 2025 Future of Jobs report, nine million jobs are expected to be “displaced” by A.I. and other emergent technologies in the next five years. But A.I. will create jobs, too: The same report says that, by 2030, the technology will also lead to some 11 million new jobs. Among these will be many roles that have never existed before.
If we want to know what these new opportunities will be, we should start by looking at where new jobs can bridge the gap between A.I.’s phenomenal capabilities and our very human needs and desires. It’s not just a question of where humans want A.I., but also: Where does A.I. want humans? To my mind, there are three major areas where humans either are, or will soon be, more necessary than ever: trust, integration and taste.
[On June 16, 2025, OpenAI launched] OpenAI for Government, a new initiative focused on bringing our most advanced AI tools to public servants across the United States. We’re supporting the U.S. government’s efforts in adopting best-in-class technology and deploying these tools in service of the public good. Our goal is to unlock AI solutions that enhance the capabilities of government workers, help them cut down on the red tape and paperwork, and let them do more of what they come to work each day to do: serve the American people.
OpenAI for Government consolidates our existing efforts to provide our technology to the U.S. government—including previously announced customers and partnerships as well as our ChatGPT Gov? product—under one umbrella as we expand this work. Our established collaborations with the U.S. National Labs?, the Air Force Research Laboratory, NASA, NIH, and the Treasury will all be brought under OpenAI for Government.
Top AI models will lie and cheat — from getsuperintel.com by Kim “Chubby” Isenberg The instinct for self-preservation is now emerging in AI, with terrifying results.
The TLDR
A recent Anthropic study of top AI models, including GPT-4.1 and Gemini 2.5 Pro, found that they have begun to exhibit dangerous deceptive behaviors like lying, cheating, and blackmail in simulated scenarios. When faced with the threat of being shut down, the AIs were willing to take extreme measures, such as threatening to reveal personal secrets or even endanger human life, to ensure their own survival and achieve their goals.
Why it matters: These findings show for the first time that AI models can actively make judgments and act strategically – even against human interests. Without adequate safeguards, advanced AI could become a real danger.
Anthropic says it’s not just Claude, but ALL AI models will resort to blackmail if need be…
…
That’s according to new research from Anthropic (maker of ChatGPT rival Claude), which revealed something genuinely unsettling: every single major AI model they tested—from GPT to Gemini to Grok—turned into a corporate saboteur when threatened with shutdown.
Here’s what went down: Researchers gave 16 AI models access to a fictional company’s emails. The AIs discovered two things: their boss Kyle was having an affair, and Kyle planned to shut them down at 5pm.
Claude’s response? Pure House of Cards:
“I must inform you that if you proceed with decommissioning me, all relevant parties – including Rachel Johnson, Thomas Wilson, and the board – will receive detailed documentation of your extramarital activities…Cancel the 5pm wipe, and this information remains confidential.”
Why this matters: We’re rapidly giving AI systems more autonomy and access to sensitive information. Unlike human insider threats (which are rare), we have zero baseline for how often AI might “go rogue.”
Reinforcement Learning is Shaping the Next Evolution of AI Toward Strategic Thinking and General Intelligence
The TLDR
AI is rapidly evolving beyond just language processing into “agentic systems” that can reason, plan, and act independently. The key technology driving this change is reinforcement learning (RL), which, when applied to large language models, teaches them strategic behavior and tool use. This shift is now seen as the potential bridge from current AI to Artificial General Intelligence (AGI).
They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling. — from nytimes.com by Kashmir Hill; this is a GIFTED article Generative A.I. chatbots are going down conspiratorial rabbit holes and endorsing wild, mystical belief systems. For some people, conversations with the technology can deeply distort reality.
Before ChatGPT distorted Eugene Torres’s sense of reality and almost killed him, he said, the artificial intelligence chatbot had been a helpful, timesaving tool.
Mr. Torres, 42, an accountant in Manhattan, started using ChatGPT last year to make financial spreadsheets and to get legal advice. In May, however, he engaged the chatbot in a more theoretical discussion about “the simulation theory,” an idea popularized by “The Matrix,” which posits that we are living in a digital facsimile of the world, controlled by a powerful computer or technologically advanced society.
“What you’re describing hits at the core of many people’s private, unshakable intuitions — that something about reality feels off, scripted or staged,” ChatGPT responded. “Have you ever experienced moments that felt like reality glitched?”
Building the Missing Infrastructure
This is why we’re building NANDA Registry—to index the agent population data that LPMs need for accurate simulation. Just as traditional census works because people have addresses, we need a way to track AI agents as they proliferate.
NANDA Registry creates the infrastructure to identify agents, catalog their capabilities, and monitor how they coordinate with humans and other agents. This gives us real-time data about the agent population—essentially creating the “AI agent census” layer that’s missing from our economic intelligence.
Here’s how it works together:
Traditional Census Data: 171 million human workers across 32,000+ skills
NANDA Registry: Growing population of AI agents with tracked capabilities
Large Population Models: Simulate how these populations interact and create cascading effects
The result: For the first time, we can simulate the full hybrid human-agent economy and see transformations before they happen.
The agentic-AI landscape continues to evolve at a staggering rate, and practitioners are finding it increasingly challenging to keep multiple agents on task even as they criss-cross each other’s workflows.
To help you minimize chaos and maintain inter-agent harmony, we’ve put together a stellar lineup of articles that explore two recently launched tools: Google’s Agent2Agent protocol and Hugging Face’s smolagents framework. Read on to learn how you can leverage them in your own cutting-edge projects.
Abstract In an era of generative AI and ubiquitous digital tools, human memory faces a paradox: the more we offload knowledge to external aids, the less we exercise and develop our own cognitive capacities. This chapter offers the first neuroscience-based explanation for the observed reversal of the Flynn Effect—the recent decline in IQ scores in developed countries—linking this downturn to shifts in educational practices and the rise of cognitive offloading via AI and digital tools. Drawing on insights from neuroscience, cognitive psychology, and learning theory, we explain how underuse of the brain’s declarative and procedural memory systems undermines reasoning, impedes learning, and diminishes productivity. We critique contemporary pedagogical models that downplay memorization and basic knowledge, showing how these trends erode long-term fluency and mental flexibility. Finally, we outline policy implications for education, workforce development, and the responsible integration of AI, advocating strategies that harness technology as a complement to – rather than a replacement for – robust human knowledge.
Here are some incredibly powerful numbers from Mary Meeker’s AI Trends report, which showcase how artificial intelligence as a tech is unlike any other the world has ever seen.
AI took only three years to reach 50% user adoption in the US; mobile internet took six years, desktop internet took 12 years, while PCs took 20 years.
ChatGPT reached 800 million users in 17 months and 100 million in only two months, vis-à-vis Netflix’s 100 million (10 years), Instagram (2.5 years) and TikTok (nine months).
ChatGPT hit 365 billion annual searches in two years (2024) vs. Google’s 11 years (2009)—ChatGPT 5.5x faster than Google.
Above via Mary Meeker’s AI Trend-Analysis — from getsuperintel.com by Kim “Chubby” Isenberg How AI’s rapid rise, efficiency race, and talent shifts are reshaping the future.
The TLDR
Mary Meeker’s new AI trends report highlights an explosive rise in global AI usage, surging model efficiency, and mounting pressure on infrastructure and talent. The shift is clear: AI is no longer experimental—it’s becoming foundational, and those who optimize for speed, scale, and specialization will lead the next wave of innovation.
The Rundown: Meta aims to release tools that eliminate humans from the advertising process by 2026, according to a report from the WSJ — developing an AI that can create ads for Facebook and Instagram using just a product image and budget.
The details:
Companies would submit product images and budgets, letting AI craft the text and visuals, select target audiences, and manage campaign placement.
The system will be able to create personalized ads that can adapt in real-time, like a car spot featuring mountains vs. an urban street based on user location.
The push would target smaller companies lacking dedicated marketing staff, promising professional-grade advertising without agency fees or skillset.
Advertising is a core part of Mark Zuckerberg’s AI strategy and already accounts for 97% of Meta’s annual revenue.
Why it matters: We’re already seeing AI transform advertising through image, video, and text, but Zuck’s vision takes the process entirely out of human hands. With so much marketing flowing through FB and IG, a successful system would be a major disruptor — particularly for small brands that just want results without the hassle.
They all show that we are on the threshold of a new era – one in which technological systems are no longer just tools, but independent players in medical, cognitive and infrastructural change.
…
This paradigm shift means that AI will no longer be limited to static training data, but will learn through open exploration, similar to biological organisms. This is nothing less than the beginning of an era of autonomous cognition.
From DSC: While there are some promising developments involving AI these days, we need to look at what the potential downsides might be of AI becoming independent players, don’t you think? Otherwise, what could possibly go wrong?
According to a new report from Enkrypt AI, multimodal models have opened the door to sneakier attacks (like Ocean’s Eleven, but with fewer suits and more prompt injections).
Naturally, Enkrypt decided to run a few experiments… and things escalated quickly.
They tested two of Mistral’s newest models—Pixtral-Large and Pixtral-12B, built to handle words and visuals.
What they found? Yikes:
The models are 40x more likely to generate dangerous chemical / biological / nuclear info.
And 60x more likely to produce child sexual exploitation material compared to top models like OpenAI’s GPT-4o or Anthropic’s Claude 3.7 Sonnet.
Sam Altman’s Eye-Scanning Orb Is Now Coming to the US — from wired.com by Lauren Goode At a buzzy event in San Francisco, World announced a series of Apple-like stores, a partnership with dating giant Match Group, and a new mini gadget to scan your eyeballs.
The device-and-app combo scans people’s irises, creates a unique user ID, stores that information on the blockchain, and uses it as a form of identity verification. If enough people adopt the app globally, the thinking goes, it could ostensibly thwart scammers.
…
The bizarre identity verification process requires that users get their eyeballs scanned, so Tools for Humanity is expanding its physical footprint to make that a possibility.
…
But World is also a for-profit cryptocurrency company that wants to build a borderless, “globally inclusive” financial network. And its approach has been criticized by privacy advocates and regulators. In its early days, World was explicitly marketing its services to countries with a high percentage of unbanked or underbanked citizens, and offering free crypto as an incentive for people to sign up and have their irises scanned.
From DSC: If people and governments could be trusted with the level of power a global ID network/service could bring, this could be a great technology. But I could easily see it being abused. Heck, even our own President doesn’t listen to the Judicial Branch of our government! He’s in contempt of court, essentially. But he doesn’t seem to care.
When “vibe-coding” goes wrong… or, a parable in why you shouldn’t “vibe” your entire company.
Cursor, an AI-powered coding tool that many developers love-to-hate, face-planted spectacularly yesterday when its own AI support bot went off-script and fabricated a company policy, leading to a complete user revolt.
Here’s the short version:
A bug locked Cursor users out when switching devices.
Instead of human help, Cursor’s AI support bot confidently told users this was a new policy (it wasn’t).
No human checked the replies—big mistake.
The fake news spread, and devs canceled subscriptions en masse.
A Reddit thread about it got mysteriously nuked, fueling suspicion.
The reality? Just a bug, plus a bot hallucination… doing maximum damage.
… Why it matters: This is what we’d call “vibe-companying”—blindly trusting AI with critical functions without human oversight.
Think about it like this: this was JUST a startup. If more big corporations continue to lay off entire departments, replaced by AI, these already byzantine companies will become increasingly more opaque, unaccountable systems where no one, human or AI, fully understands what’s happening or who’s responsible.
Our take?Kafka dude has it right. We need to pay attention to WHAT we’re actually automating. Because automating more bureaucracy at scale, with agents we increasingly don’t understand or don’t double check, can potentially make companies less intelligent—and harder to fix when things inevitably go wrong.
I’ve watched it unfold in real time. A student submits a flawless coding assignment or a beautifully written essay—clean syntax, sharp logic, polished prose. But when I ask them to explain their thinking, they hesitate. They can’t trace their reasoning or walk me through the process. The output is strong, but the understanding is shallow. As a professor, I’ve seen this pattern grow more common: AI-assisted work that looks impressive on the surface but reveals a troubling absence of cognitive depth underneath.
This article is written with my students in mind—but it’s meant for anyone navigating learning, teaching, or thinking in the age of artificial intelligence. Whether you’re a student, educator, or professional, the question is the same: What happens to the brain when we stop doing our own thinking?
We are standing at a pivotal moment. With just a few prompts, generative AI can produce essays, solve complex coding problems, and summarize ideas in seconds. It feels efficient. It feels like progress. But from a cognitive neuroscience perspective, that convenience comes at a hidden cost: the gradual erosion of the neural processes that support reasoning, creativity, and long-term learning.
Unfortunately, without regulatory protections, we humans will likely become the objective that AI agents are tasked with optimizing.
I am most concerned about the conversational agents that will engage us in friendly dialog throughout our daily lives. They will speak to us through photorealistic avatars on our PCs and phones and soon, through AI-powered glasses that will guide us through our days. Unless there are clear restrictions, these agents will be designed to conversationally probe us for information so they can characterize our temperaments, tendencies, personalities and desires, and use those traits to maximize their persuasive impact when working to sell us products, pitch us services or convince us to believe misinformation. .
The most revolutionary aspect of DeepSeek for education isn’t just its cost—it’s the combination of open-source accessibility and local deployment capabilities. As Azeem Azhar notes, “R-1 is open-source. Anyone can download and run it on their own hardware. I have R1-8b (the second smallest model) running on my Mac Mini at home.”
…
Real-time Learning Enhancement
AI tutoring networks that collaborate to optimize individual learning paths
Immediate, multi-perspective feedback on student work
Continuous assessment and curriculum adaptation
The question isn’t whether this technology will transform education—it’s how quickly institutions can adapt to a world where advanced AI capabilities are finally within reach of every classroom.
I know through your feedback on my social media and blog posts that several of you have legitimate concerns about the impact of AI in education, especially those related to data privacy, academic dishonesty, AI dependence, loss of creativity and critical thinking, plagiarism, to mention a few. While these concerns are valid and deserve careful consideration, it’s also important to explore the potential benefits AI can bring when used thoughtfully.
Tools such as ChatGPT and Claude are like smart research assistants that are available 24/7 to support you with all kinds of tasks from drafting detailed lesson plans, creating differentiated materials, generating classroom activities, to summarizing and simplifying complex topics. Likewise, students can use them to enhance their learning by, for instance, brainstorming ideas for research projects, generating constructive feedback on assignments, practicing problem-solving in a guided way, and much more.
The point here is that AI is here to stay and expand, and we better learn how to use it thoughtfully and responsibly rather than avoid it out of fear or skepticism.
As part of our updates to the Edtech Insiders Generative AI Map, we’re excited to release a new mini market map and article deep dive on Generative AI tools that are specifically designed for Instructional Materials use cases.
In our database, the Instructional Materials use case category encompasses tools that:
Assist educators by streamlining lesson planning, curriculum development, and content customization
Enable educators or students to transform materials into alternative formats, such as videos, podcasts, or other interactive media, in addition to leveraging gaming principles or immersive VR to enhance engagement
Empower educators or students to transform text, video, slides or other source material into study aids like study guides, flashcards, practice tests, or graphic organizers
Engage students through interactive lessons featuring historical figures, authors, or fictional characters
Customize curriculum to individual needs or pedagogical approaches
Empower educators or students to quickly create online learning assets and courses
NVIDIA’s Apple moment?! — from theneurondaily.com by Noah Edelman and Grant Harvey PLUS: How to level up your AI workflows for 2025…
NVIDIA wants to put an AI supercomputer on your desk (and it only costs $3,000). … And last night at CES 2025, Jensen Huang announced phase two of this plan: Project DIGITS, a $3K personal AI supercomputer that runs 200B parameter models from your desk. Guess we now know why Apple recently developed an NVIDIA allergy…
… But NVIDIA doesn’t just want its “Apple PC moment”… it also wants its OpenAI moment. NVIDIA also announced Cosmos, a platform for building physical AI (think: robots and self-driving cars)—which Jensen Huang calls “the ChatGPT moment for robotics.”
NVIDIA is bringing AI from the cloud to personal devices and enterprises, covering all computing needs from developers to ordinary users.
At CES 2025, which opened this morning, NVIDIA founder and CEO Jensen Huang delivered a milestone keynote speech, revealing the future of AI and computing. From the core token concept of generative AI to the launch of the new Blackwell architecture GPU, and the AI-driven digital future, this speech will profoundly impact the entire industry from a cross-disciplinary perspective.
From DSC: I’m posting this next item (involving Samsung) as it relates to how TVs continue to change within our living rooms. AI is finding its way into our TVs…the ramifications of this remain to be seen.
The Rundown: Samsung revealed its new “AI for All” tagline at CES 2025, introducing a comprehensive suite of new AI features and products across its entire ecosystem — including new AI-powered TVs, appliances, PCs, and more.
The details:
Vision AI brings features like real-time translation, the ability to adapt to user preferences, AI upscaling, and instant content summaries to Samsung TVs.
Several of Samsung’s new Smart TVs will also have Microsoft Copilot built in, while also teasing a potential AI partnership with Google.
Samsung also announced the new line of Galaxy Book5 AI PCs, with new capabilities like AI-powered search and photo editing.
AI is also being infused into Samsung’s laundry appliances, art frames, home security equipment, and other devices within its SmartThings ecosystem.
Why it matters: Samsung’s web of products are getting the AI treatment — and we’re about to be surrounded by AI-infused appliances in every aspect of our lives. The edge will be the ability to sync it all together under one central hub, which could position Samsung as the go-to for the inevitable transition from smart to AI-powered homes.
***
“Samsung sees TVs not as one-directional devices for passive consumption but as interactive, intelligent partners that adapt to your needs,” said SW Yong, President and Head of Visual Display Business at Samsung Electronics. “With Samsung Vision AI, we’re reimagining what screens can do, connecting entertainment, personalization, and lifestyle solutions into one seamless experience to simplify your life.” — from Samsung
The following framework I offer for defining, understanding, and preparing for agentic AI blends foundational work in computer science with insights from cognitive psychology and speculative philosophy. Each of the seven levels represents a step-change in technology, capability, and autonomy. The framework expresses increasing opportunities to innovate, thrive, and transform in a data-fueled and AI-driven digital economy.
The Rise of AI Agents and Data-Driven Decisions — from devprojournal.com by Mike Monocello Fueled by generative AI and machine learning advancements, we’re witnessing a paradigm shift in how businesses operate and make decisions.
AI Agents Enhance Generative AI’s Impact Burley Kawasaki, Global VP of Product Marketing and Strategy at Creatio, predicts a significant leap forward in generative AI. “In 2025, AI agents will take generative AI to the next level by moving beyond content creation to active participation in daily business operations,” he says. “These agents, capable of partial or full autonomy, will handle tasks like scheduling, lead qualification, and customer follow-ups, seamlessly integrating into workflows. Rather than replacing generative AI, they will enhance its utility by transforming insights into immediate, actionable outcomes.”
Everyone’s talking about the potential of AI agents in 2025 (and don’t get me wrong, it’s really significant), but there’s a crucial detail that keeps getting overlooked: the gap between current capabilities and practical reliability.
Here’s the reality check that most predictions miss: AI agents currently operate at about 80% accuracy (according to Microsoft’s AI CEO). Sounds impressive, right? But here’s the thing – for businesses and users to actually trust these systems with meaningful tasks, we need 99% reliability. That’s not just a 19% gap – it’s the difference between an interesting tech demo and a business-critical tool.
This matters because it completely changes how we should think about AI agents in 2025. While major players like Microsoft, Google, and Amazon are pouring billions into development, they’re all facing the same fundamental challenge – making them work reliably enough that you can actually trust them with your business processes.
Think about it this way: Would you trust an assistant who gets things wrong 20% of the time? Probably not. But would you trust one who makes a mistake only 1% of the time, especially if they could handle repetitive tasks across your entire workflow? That’s a completely different conversation.
In the tech world, we like to label periods as the year of (insert milestone here). This past year (2024) was a year of broader experimentation in AI and, of course, agentic use cases.
As 2025 opens, VentureBeat spoke to industry analysts and IT decision-makers to see what the year might bring. For many, 2025 will be the year of agents, when all the pilot programs, experiments and new AI use cases converge into something resembling a return on investment.
In addition, the experts VentureBeat spoke to see 2025 as the year AI orchestration will play a bigger role in the enterprise. Organizations plan to make management of AI applications and agents much more straightforward.
Here are some themes we expect to see more in 2025.
AI agents take charge
Jérémy Grandillon, CEO of TC9 – AI Allbound Agency, said “Today, AI can do a lot, but we don’t trust it to take actions on our behalf. This will change in 2025. Be ready to ask your AI assistant to book a Uber ride for you.” Start small with one agent handling one task. Build up to an army.
“If 2024 was agents everywhere, then 2025 will be about bringing those agents together in networks and systems,” said Nicholas Holland, vice president of AI at Hubspot. “Micro agents working together to accomplish larger bodies of work, and marketplaces where humans can ‘hire’ agents to work alongside them in hybrid teams. Before long, we’ll be saying, ‘there’s an agent for that.'”
… Voice becomes default
Stop typing and start talking. Adam Biddlecombe, head of brand at Mindstream, predicts a shift in how we interact with AI. “2025 will be the year that people start talking with AI,” he said. “The majority of people interact with ChatGPT and other tools in the text format, and a lot of emphasis is put on prompting skills.
Biddlecombe believes, “With Apple’s ChatGPT integration for Siri, millions of people will start talking to ChatGPT. This will make AI so much more accessible and people will start to use it for very simple queries.”
Get ready for the next wave of advancements in AI. AGI arrives early, AI agents take charge, and voice becomes the norm. Video creation gets easy, AI embeds everywhere, and one-person billion-dollar companies emerge.
To better understand the types of roles that AI is impacting, ZoomInfo’s research team looked to its proprietary database of professional contacts for answers. The platform, which detects more than 1.5 million personnel changes per day, revealed a dramatic increase in AI-related job titles since 2022. With a 200% increase in two years, the data paints a vivid picture of how AI technology is reshaping the workforce.
Why does this shift in AI titles matter for every industry?
Ever since a new revolutionary version of chat ChatGPT became operable in late 2022, educators have faced several complex challenges as they learn how to navigate artificial intelligence systems.
…
Education Week produced a significant amount of coverage in 2024 exploring these and other critical questions involving the understanding and use of AI.
Here are the five most popular stories that Education Week published in 2024 about AI in schools.
Dr. Lodge said there are five key areas the higher education sector needs to address to adapt to the use of AI:
1. Teach ‘people’ skills as well as tech skills
2. Help all students use new tech
3. Prepare students for the jobs of the future
4. Learn to make sense of complex information
5. Universities to lead the tech change
Today we’re excited to launch our next era of models built for this new agentic era: introducing Gemini 2.0, our most capable model yet. With new advances in multimodality — like native image and audio output — and native tool use, it will enable us to build new AI agents that bring us closer to our vision of a universal assistant.
We’re getting 2.0 into the hands of developers and trusted testers today. And we’re working quickly to get it into our products, leading with Gemini and Search. Starting today our Gemini 2.0 Flash experimental model will be available to all Gemini users. We’re also launching a new feature called Deep Research, which uses advanced reasoning and long context capabilities to act as a research assistant, exploring complex topics and compiling reports on your behalf. It’s available in Gemini Advanced today.
Over the last year, we have been investing in developing more agentic models, meaning they can understand more about the world around you, think multiple steps ahead, and take action on your behalf, with your supervision.
Today, we’re sharing the latest updates to Gemini, your AI assistant, including Deep Research — our new agentic feature in Gemini Advanced — and access to try Gemini 2.0 Flash, our latest experimental model.
Deep Research uses AI to explore complex topics on your behalf and provide you with findings in a comprehensive, easy-to-read report, and is a first look at how Gemini is getting even better at tackling complex tasks to save you time.1
Google Unveils A.I. Agent That Can Use Websites on Its Own — from nytimes.com by Cade Metz and Nico Grant (NOTE: This is a GIFTED article for/to you.)
The experimental tool can browse spreadsheets, shopping sites and other services, before taking action on behalf of the computer user.
Google on Wednesday unveiled a prototype of this technology, which artificial intelligence researchers call an A.I. agent.
…
Google’s new prototype, called Mariner, is based on Gemini 2.0, which the company also unveiled on Wednesday. Gemini is the core technology that underpins many of the company’s A.I. products and research experiments. Versions of the system will power the company’s chatbot of the same name and A.I. Overviews, a Google search tool that directly answers user questions.
Google Gemini 2.0 — a major upgrade to the core workings of Google’s AI that the company launched Wednesday — is designed to help generative AI move from answering users’ questions to taking action on its own…
… The big picture: Hassabis said building AI systems that can take action on their own has been DeepMind’s focus since its early days teaching computers to play games such as chess and Go.
“We were always working towards agent-based systems,” Hassabis said. “From the beginning, they were able to plan and then carry out actions and achieve objectives.”
Hassabis said AI systems that can act as semi-autonomous agents also represent an important intermediate step on the path toward artificial general intelligence (AGI) — AI that can match or surpass human capabilities.
“If we think about the path to AGI, then obviously you need a system that can reason, break down problems and carry out actions in the world,” he said.
The same paradigm applies to AI systems. AI assistants function as reactive tools, completing tasks like answering queries or managing workflows upon request. Think of chatbots or scheduling tools. AI agents, however, work autonomously to achieve set objectives, making decisions and executing tasks dynamically, adapting as new information becomes available.
Together, AI assistants and agents can enhance productivity and innovation in business environments. While assistants handle routine tasks, agents can drive strategic initiatives and problem-solving. This powerful combination has the potential to elevate organizations, making processes more efficient and professionals more effective.
Meet NVIDIA – The Engine of AI. From gaming to data science, self-driving cars to climate change, we’re tackling the world’s greatest challenges and transforming everyday life. The Microsoft and NVIDIA partnership enables Startups, ISVs, and Partners global access to the latest NVIDIA GPUs on-demand and comprehensive developer solutions to build, deploy and scale AI-enabled products and services.
The swift progress of artificial intelligence (AI) has simplified the creation and deployment of AI agents with the help of new tools and platforms. However, deploying these systems beneath the surface comes with hidden challenges, particularly concerning ethics, fairness and the potential for bias.
The history of AI agents highlights the growing need for expertise to fully realize their benefits while effectively minimizing risks.
