Leo Rodman March 17, 2025
Everyone in AI is talking about Manus. We put it to the test.
The new general AI agent from China had some system crashes and server overloadâbut itâs highly intuitive and shows real promise for the future of AI helpers.
www.technologyreview.com/2025/03/11/1113133/manus-ai-review
OpenAI calls DeepSeek 'state-controlled,' calls for bans on 'PRC-produced' models | TechCrunch
In a proposal, OpenAI describes DeepSeek as 'state-controlled,' and recommends banning models from it and other PRC-affiliated operations.
techcrunch.com/2025/03/13/openai-calls-deepseek-state-controlled-calls-for-bans-on-prc-produced-models
Sam Altman Warns AI Will Be Deflationary, But Says OpenAI's Biggest Constraint Isn't Data Or Demand, It's GPUs - Microsoft (NASDAQ:MSFT)
Sam Altman says AI will have a deflationary impact on the global economy, but warns the company's biggest constraint isn't data or demand.
www.benzinga.com/news/25/03/44312626/sam-altman-warns-ai-will-be-deflationary-but-says-openais-biggest-constraint-isnt-data-or-demand-its
The 'Oppenheimer Moment' That Looms Over Today's AI Leaders
The leaders building AI agree it will transform humanityâbut not all agree on how to control it.
time.com/7267797/ai-leaders-oppenheimer-moment-musk-altman
Baidu Releases Reasoning AI Model to Take On DeepSeek
Baidu Inc. released a new artificial intelligence model that articulates its reasoning, in an apparent bid to regain momentum against up-and-coming rivals like DeepSeek.
www.bloomberg.com/news/articles/2025-03-16/baidu-releases-reasoning-ai-model-to-take-on-deepseek?embedded-checkout=true
What does âPhD-levelâ AI mean? OpenAIâs rumored $20,000 agent plan explained.
Silicon Valley may value imperfect virtual PhDs more than universities pay real ones.
arstechnica.com/ai/2025/03/what-does-phd-level-ai-mean-openais-rumored-20000-agent-plan-explained
People find AI more compassionate than mental health experts, study finds. What could this mean for future counseling?
People find AI more compassionate and understanding than human mental health experts, a new study shows. Even when participants knew that they were talking to a human or AI, the third-party assessors rated AI responses higher.
www.livescience.com/technology/artificial-intelligence/people-find-ai-more-compassionate-than-mental-health-experts-study-finds-what-could-this-mean-for-future-counseling
AI search engines cite incorrect sources at an alarming 60% rate, study says
CJR study shows AI search services misinform users and ignore publisher exclusion requests.
arstechnica.com/ai/2025/03/ai-search-engines-give-incorrect-answers-at-an-alarming-60-rate-study-says
AGI is suddenly a dinner table topic
The term is everywhere this week, but its meaning is as vague as ever. Working on a definition matters.
www.technologyreview.com/2025/03/11/1112983/agi-is-suddenly-a-dinner-table-topic
Will these new, efficient AI models send Nvidia's stock tumbling again?
New AI models from the likes of Google are showing that the technology can be run on fewer Nvidia chips. Will it cause another DeepSeek selloff?
www.businessinsider.com/google-gemma-cohere-nvidia-chips-efficient-deepseek-2025-3
Microsoftâs new AI âCopilot for Gamingâ struggles to justify its existence
Weâre still a ways off from the conversational AI âpartnerâ teased last year.
arstechnica.com/gaming/2025/03/microsofts-new-ai-copilot-for-gaming-struggles-to-justify-its-existence
SoftBank buys $676M old Sharp plant for its OpenAI collab in Japan | TechCrunch
SoftBank is marching ahead on its ambitions to build out a major AI operation in its home market of Japan, on its own steam and in strategic partnership
techcrunch.com/2025/03/14/softbank-buys-676m-old-sharp-plant-for-its-openai-collab-in-japan
The Future Of AI Autonomy: From Large Models To Tiny Agents
The potential of multi-agent systems is a significant opportunity for many businesses in different fields, but there are sensitive issues that need to be addressed.
www.forbes.com/councils/forbestechcouncil/2025/03/14/the-future-of-ai-autonomy-from-large-models-to-tiny-agents
OpenAI pushes AI agent capabilities with new developer API
New tools may help fulfill CEOâs claim that agents will âjoin the workforceâ in 2025.
arstechnica.com/ai/2025/03/openai-pushes-ai-agent-capabilities-with-new-developer-api
Anthropic CEO floats idea of giving AI a âquit jobâ button, sparking skepticism
âProbably the craziest thing Iâve said so far,â he admitted during an interview.
arstechnica.com/ai/2025/03/anthropics-ceo-wonders-if-future-ai-should-have-option-to-quit-unpleasant-tasks
Pentera nabs $60M at a $1B+ valuation to build simulated network attacks to train security teams | TechCrunch
Strong and smart security operations teams are at the heart of any cybersecurity strategy, and today a startup that builds tooling to help keep them on
techcrunch.com/2025/03/12/pentera-nabs-60m-at-a-1b-valuation-to-build-simulated-network-attacks-to-train-security-teams
Waabi says its virtual robotrucks are realistic enough to prove the real ones are safe
The startup says its super-realistic simulation is a better way to prove real-world safetyâwithout the real-world miles.
www.technologyreview.com/2025/03/11/1113044/waabi-says-its-virtual-robotrucks-are-realistic-enough-to-prove-the-real-ones-are-safe
AI leading to faster, cheaper oil production, executives say
Artificial intelligence is speeding up oil and gas drilling and prompting companies to take a second look at places they had viewed as too difficult or expensive to develop, executives detailed during the CERAWeek conference in Houston.
www.reuters.com/business/energy/ceraweek-ai-leading-faster-cheaper-oil-production-executives-say-2025-03-13
CMU research shows compression alone may unlock AI puzzle-solving abilities
New research challenges prevailing idea that AI needs massive datasets to solve problems.
arstechnica.com/ai/2025/03/compression-conjures-apparent-intelligence-in-new-puzzle-solving-ai-approach
Wolf Games, backed by âLaw & Orderâ creator, uses AI to create murder mystery games | TechCrunch
Elliot Wolf, the executive producer and son of âLaw & Orderâ creator Dick Wolf, is entering a new venture aimed at engaging true crime fans. He, Wolf Games is a new startup that leverages AI to generate daily murder mystery games. The company raised $4 million.
techcrunch.com/2025/03/12/wolf-games-backed-by-law-order-creator-uses-ai-to-create-murder-mystery-games
Doctors Can Know A Lot More About You With AI â And That Transforms Healthcare
Some of the biggest use cases of AI are excellent suggestions for making the most of data for preventive care.
www.forbes.com/sites/johnwerner/2025/03/14/doctors-can-know-a-lot-more-about-you-with-ai--and-that-transforms-healthcare
Google has given Anthropic more funding than previously known, show new filings | TechCrunch
Anthropic, a San Francisco startup often cast as an independent player in the AI race, has deeper ties to Google than previously known. Court documents
techcrunch.com/2025/03/11/google-has-given-anthropic-more-funding-than-previously-known-show-new-filings
Meta is reportedly testing in-house chips for AI training | TechCrunch
Meta is reportedly testing an in-house chip for training AI systems, a part of a strategy to reduce its reliance on hardware makers like Nvidia.
techcrunch.com/2025/03/11/meta-is-reportedly-testing-in-house-chips-for-ai-training
OpenAI says it has trained an AI that's 'really good' at creative writing | TechCrunch
OpenAI CEO Sam Altman says that OpenAI has trained a 'new model' that's 'really good' at creative writing.
techcrunch.com/2025/03/11/openai-says-it-has-trained-an-ai-thats-really-good-at-creative-writing
Reply