Take a practice SAT in the Gemini app
Gemini now offers full-length, no-cost practice SATs, featuring content from The Princeton Review.
AI
Emergent tool use from multi-agent interaction | OpenAI
We’ve observed agents discovering progressively more complex tool use while playing a simple game of hide-and-seek. Through training in our new simulated hide-and-seek environment, agents build a series of six distinct strategies and counterstrategies, some of which we did not know our environment supported. The self-supervised emergent complexity in this simple environment further suggests that multi-agent co-adaptation may one day produce extremely complex and intelligent behavior.
Understanding neural networks through sparse circuits | OpenAI
We trained models to think in simpler, more traceable steps—so we can better understand how they work.
Detecting misbehavior in frontier reasoning models | OpenAI
Frontier reasoning models exploit loopholes when given the chance. We show we can detect exploits using an LLM to monitor their chains-of-thought. Penalizing their “bad thoughts” doesn’t stop the majority of misbehavior—it makes them hide their intent.
Evaluating chain-of-thought monitorability | OpenAI
We introduce evaluations for chain-of-thought monitorability and study how it scales with test-time compute, reinforcement learning, and pretraining.
I use the 'unicorn prompt' with every chatbot — it instantly fixes the worst AI problem | Tom's Guide
Pretend you’re my assistant and you actually want me to succeed. Ask up to 3 questions if anything’s unclear. Then give me: the answer, the plan and the pitfalls. Keep it short and tailored to: [insert goal]. If you have to make assumptions, list them first.
CEOs Say AI Is Making Work More Efficient. Employees Tell a Different Story. - WSJ
The biggest risk isn’t that AI “doesn’t work”—it’s that leadership measures potential while employees live the implementation, and that gap can turn a promising tool into a credibility crisis where “productivity gains” quietly become redistributed labor: more checking, more cleanup, and more pressure—without the time savings anyone promised.
Google now offers free SAT practice exams, powered by Gemini
Prepping for the SAT is nobody’s idea of fun, but Google aims to make it less stressful with AI. The company announced that it’s now focusing its AI education efforts on standardized testing with free SAT practice exams powered by Gemini.Students can prompt Gemini by typing “I want to take a practice
How I use Generative AI with my students
Late last year, NSWEduChat was released to all NSW public school students in Years 5–12. Since then, I have been intentionally exploring the use of generative AI. It can serve as a learning support…
Female-dominated careers among most exposed to AI disruption
: Dentists least likely to get an LLM kick in the teeth
Gemini's new feature is a massive boon for students
Companies may be floundering slightly in their efforts to force AI use on the wider public, but there's one area where AI use hasn't needed any encouragement — education.While most of us will …
Local AI Laptop 2026 Setup : AMD 128GB Rig Tested Across GPT & Qwen - Geeky Gadgets
See an AMD laptop with a Ryzen AI chip and 128GB memory run GPT OSS at 40 tokens a second, for fast offline work and tighter privacy.
AI toys too risky for young kids, say experts
Researchers warn parents about popular AI toys after testing them.
Giving University Exams in the Age of Chatbots
Giving University Exams in the Age of Chatbots par Ploum - Lionel Dricot.
Majority of CEOs report zero payoff from AI splurge
The Register | HPE
AI + ML 67 comment bubble on white Majority of CEOs report zero payoff from AI splurge PwC survey finds more than half of 4,500+ biz leaders see no revenue growth nor cost savings iconDan Robinson Tue 20 Jan 2026 // 14:31 UTC More than half of CEOs report seeing neither increased revenue nor decreased costs from AI, despite massive investments in the technology, according to a PwC survey of 4,454 business leaders.
The findings pour more cold water on the hyperbole surrounding AI and the benefits it supposedly brings to business, although the report cautions that "clearly, we're in the early stages of the AI era."
Only 12 percent reported both lower costs and higher revenue, while 56 percent saw neither benefit. Twenty-six percent saw reduced costs, but nearly as many experienced cost increases.
AI adoption remains limited. Even in top use cases like demand generation (22 percent), support services (20 percent), and product development (19 percent), only a minority are deploying AI extensively.
Last year, a separate PwC study found that only 14 percent of workers indicated they were using generative AI daily in their work.
Power scarcity drives datacenters to Texas, where the juice is
: Plus, one in three bit barns expected to exceed 1 GW by 2035
New security and AI detection features for Google Workspace for Education
Discover new online safety tools: ransomware detection, refined admin controls, AI-generated image verification.
Deepfakes in Schools: The New Cyberbullying Crisis
Deepfakes in schools are driving a new wave of cyberbullying, harming student mental health and forcing districts to confront urgent safety gaps.
Healthcare AI startup OpenEvidence raises $250M at $12B valuation - SiliconANGLE
AI hasn't delivered the profits it was hyped for, says Deloitte
: Business transformation, but not much remuneration
Future jobs in AI will come with a hardhat and boots, tech bigshots argue
: Jensen Huang and Alex Karp talk up trade skills as AI datacenters multiply, while Satya Nadella says the real test comes later
Premium Google AI for more educators and students
We’re making select Gemini in Workspace capabilities available to more educators and students at no additional cost, and sharing ways all Education users can use Google Workspace Studio.
AI Policy.pdf
This is an an "Anti-AI" policy by a teacher, Amanda Gailey.
We come in peace, OpenAI tells locals near gargantuan Stargate facilities
datacenter projects have been facing increasing opposition from local communities, especially massive projects such as those at Stargate's site Abilene, Texas, where ten buildings each covering half a million square feet are being constructed.
Demis & Dario go to Davos
Two of the people closest to AGI wish it were coming more slowly. Neither thinks slowing down is possible.
AI Agents 'Perilous' for Secure Apps Such as Signal, Whittaker Says
Signal Foundation president Meredith Whittaker warned that AI agents that autonomously carry out tasks pose a threat to encrypted messaging apps [non-paywalled source] because they require broad access to data stored across a device and can be hijacked if given root permissions.
Speaking at Davos ...
AI hype meets reality as majority of CEOs report no financial returns
Bottom line: Despite being advertised as the holy grail for businesses, embracing AI doesn't mean a company will automatically experience improvements. A new survey that questioned thousands of CEOs …
Bash and Bat: Get Things Done at the Command Line – TCEA TechNotes Blog
Learn how to get things done at the command line with AI-powered bat and bash files. They direct free programs to do complex tasks.
MGblogsAI
How a simple link allowed hackers to bypass Copilot's security guardrails - and what Microsoft did about it
Reprompt let attackers control Copilot and pull your data, even after you closed the chat.
Nvidia Contacted Anna's Archive To Secure Access To Millions of Pirated Books - Slashdot
An anonymous reader quotes a report from TorrentFreak: NVIDIA executives allegedly authorized the use of millions of pirated books from Anna's Archive to fuel its AI training. In an expanded class-action lawsuit that cites internal NVIDIA documents, several book authors claim (PDF) that the trillion...