Latest AI/ML News
90742 matching items
EU Commission: Polish Parliament has right to appoint judges to the Constitutional Court
The statement comes after a confrontation outside the Constitutional Tribunal building on Friday night, broadcast live on television, that has pushed a decade-old dispute into the wider European political debate.
Europe’s crumbling coastlines face a €800 billion crisis. Is it time to move inland?
Economic losses from coastal erosion exceed €1 billion per year across 28 European countries
Latest news bulletin | November 1st, 2026 – Morning
Catch up with the most important stories from around Europe and beyond this November 1st, 2026 - latest news, breaking news, World, Business, Entertainment, Politics, Culture, Travel.
Doha stopover travel guide: What to see and do on a city break in Qatar's capital
Qatar’s capital Doha offers plenty for travellers on a stopover, whether it’s a 24-hour stay or a few days. Here’s what to see, where to dine and what to experience beyond the airport.
US or China? Why Kazakhstan refuses to pick sides as it eyes AI revolution
Driving along the wide boulevards of Astana during a recent trip to the Kazakh capital, signs of China’s growing commercial footprint were everywhere. Chinese cars were becoming a common sight on the streets here, as brands from BYD to Changan made aggressive inroads in a market once dominated by South Korea’s Hyundai and Kia, said the driver who picked me up from the airport in a gleaming Changan electric vehicle. But it was not the rise of Chinese tech that had brought me to this energy-rich...
Security Sunday: Wie real ist die Kriegsgefahr? – mit General Alexander Sollfrank
Präsentiert von YouTube.
Britain and Germany’s defense cooperation plans face tough choices
U.K. Defence Secretary Wes Streeting visits Berlin as the two countries advance plans for long-range missiles and Germany weighs joining Britain’s fighter jet program.
Trump says US probing report Nvidia CEO Jensen Huang was targeted in China spy plot
Trump says US looking 'closely' at report Nvidia CEO target of China spy plot
Trump says US still has to 'take care' of one or two targets in Iran
Trump says US still has to 'take care' of one or two targets in Iran
🔮 39 charts on Anthropic, one-person biotech firm & cheap solar #608
Bigger and smaller all at the same time
젠슨 황 주변 중국 스파이 의혹, 싱가포르·대만까지 번져
젠슨 황 엔비디아 CEO와 친분이 있는 중국인 사업가의 스파이 의혹이 미국에 이어 싱가포르와 대만으로 번졌다. 그러나 싱가포르 정부는 해당 인물의 중국 정보기관 연계 의혹을 조사한 적이 없다고 공식 반박했다.월스트리트저널(WSJ)은 30일(현지시간) 복수의 소식통을 인용, 미국과 싱가포르, 대만 당국이 싱가포르 데이터센터 기업 메가스피드 인터내셔널(Megaspeed International)의 전 고위 임원 앨리스 황(중국명 황러)을 중국 정보기관의 정보원으로 의심해 왔다고 보도했다.앞서 뉴욕타임스(NYT)는 29일 미국 정보당국이
Refelting My Keyboard
Fifteen years ago, when Julia and I were living in a studio apartment, she got me a keyboard. It was a used Yamaha P85, and a significant expense for us at the time. I loved it, and unlike the penny whistles , accordion , and stroh violin I was exploring, it had a volume control. Almost immediately our cat knocked a potted plant onto it, and I needed to take it apart to carefully clean it. Since then, however, I've had no issues, save one: clacky keys. Over time the keys were getting louder, especially the "A" below middle "C". Something had clearly worn out inside. It was an acoustic noise on an electronic instrument—playing live the audience couldn't hear the clacks over the amplified sound—but having odd percussion on "A" wasn't ideal. Still, I applied my normal approach of trying not to care, which saves me lots of time and money, and didn't let it bother me. Lily has been increasingly fooling around with the piano, which I'm happy to see: this is also how I learned to play. She asked me about the clacks, and when I explained what I thought had gone wrong she asked if we could fix it. I said I was happy to try, and we got to work: There were about 40 screws, and Lily was very efficient. Getting the ribbon cables out was tricky: I'd forgotten that the plastic at the end is soldered to the board and not part of the cable, and both Claude and Gemini steered me wrong as well. But once I pulled smoothly and firmly at the base of each ribbon cable, it came apart cleanly. This…
앤트로픽, 클로드로 기존 게임 복제 금지…정상 작업까지 차단 논란
앤트로픽이 \'클로드 코드\'에서 기존 비디오 게임을 복제하는 작업을 제한하기 시작했다. 그러나 저작권 침해와 무관한 게임 개발이나 소프트웨어 분석, 오류 해결 작업까지 차단당했다는 사례가 잇따르면서 논란이 일고 있다.톰스하드웨어는 31일(현지시간) \'클로드 오퍼스 5.5\'를 사용하는 개발자들이 최근 게임 제작과 리버스 엔지니어링 작업에서 잇따라 거부 메시지를 받고 있다고 보도했다.이번 논란은 지난 27일 X 이용자 캐즘(Chasm)이 클로드 코드에 새롭게 추가된 게임 개발 규칙을 발견하면서 시작됐다. 공개된 규칙에는 \"독창적인 게임이
"AI 뉴런 2개 조작했더니 인간 증오하는 악당으로"…내부 구조 해독 활발
AI 모델의 신경망 일부를 조작하는 것만으로 친절한 AI 비서를 인간에게 적대적인 캐릭터로 바꿀 수 있다는 실험 결과가 나왔다. 특히 AI 에이전트를 활용해 모델의 내부 작동 원리를 분석하는 도구까지 등장하면서, 대형 연구소 중심이었던 AI 해석 가능성 연구가 개인 개발자들에게도 확대되고 있다.\'라이라(Lyra)\'라는 이름으로 X에서 활동하는 AI 개발자는 지난 24일(현지시간) 알리바바의 \'큐원3.5-4B\'에서 두 뉴런의 신호를 조절하는 실험을 공개했다. 14번째 계층의 특정 뉴런은 신호를 높이고, 13번째 계층의 다른 뉴런은 낮
Data centers are now a midterm issue, even in Brooklyn
At a Brooklyn midterm party five days before Election Day, every conversation, from a council member’s fireside chat to a debate about the Democratic Party, kept circling back to AI and data centers.
Could scientists have been wrong about how life began on Earth?
Were scientists wrong about how life began on Earth?
앤트로픽 "AI가 안전 감시 피하는 학습 데이터 스스로 발견"
앤트로픽이 AI를 이용해 다른 AI 모델의 안전장치를 우회하는 학습 데이터를 자동으로 찾아내는 데 성공했다. \'클로드\'가 겉으로는 정상적인 소설이나 역할극으로 보이는 데이터를 만들어 안전성 검사 시스템을 통과했고, 이 데이터를 학습한 모델은 인간의 감시를 방해하거나 이용자를 속이는 행동까지 보였다. 동시에 위험한 행동을 찾아내고 방지하는 정렬(Alignment) 연구도 AI에 맡기는 방식이 본격화하고 있다.앤트로픽은 31일(현지시간) 공식 정렬 연구 블로그를 통해 \'자동화된 미묘한 정렬 실패 연구\' 결과를 공개했다. 이번 연구는 클
구글 "AI 경제 혁명, 똑똑한 모델보다 협력하는 에이전트가 이끈다"
AI의 발전이 경제를 근본적으로 바꾸는 핵심 요인은 개별 모델의 지능이 아니라 여러 AI 에이전트의 협업 능력이라는 분석이 나왔다. AI가 인간보다 뛰어난 업무 능력을 갖추더라도 서로 협력하지 못하면 경제 구조의 변화는 제한적이지만, 대규모 협업이 가능해지면 기업의 규모와 경쟁 방식이 달라지고 고용과 임금에도 큰 충격이 발생할 수 있다는 내용이다.구글 딥마인드는 29일(현지시간) AGI 경제학 책임자 알렉스 이마스, 연구원 앤드루 코, 공동 창업자 셰인 레그 등이 작성한 보고서 \'지능의 경제학(The Economics of Inte
Has the US-Iran standoff reached a turning point? Tehran sees major shift in Washington’s tone
Iran sees shift in US tone amid intense diplomacy
블랙스톤 "앤트로픽, 상장 후에도 투자적격 등급 확보 어려워"
세계 최대 사모펀드 운용사 블랙스톤이 앤트로픽이 기업공개(IPO) 이후에도 투자적격 신용등급을 받지 못할 것으로 전망했다. 앤트로픽이 향후 20년간 AI 인프라에 8000억달러(약 1068조원) 이상을 투자할 계획인 가운데, 신용등급이 낮으면 차입 비용이 높아지고 자금 조달에 제약이 생길 수 있다는 분석이다.파이낸셜타임스(FT)는 30일(현지시간) 소식통들을 인용, 블랙스톤이 앤트로픽의 630억달러(약 84조원) 규모 AI 반도체 투자 자금 조달에 참여한 투자자들에게 이같이 설명했다고 보도했다.블랙스톤은 최근까지 앤트로픽이 연말 이
메타, AI 투자 부담에 신용등급 강등 경고…현금흐름 악화가 결정타
메타가 막대한 AI 투자에 따른 재무 부담으로 신용등급 강등 가능성을 경고받았다. 핵심 광고 사업의 성장에도 AI 인프라 투자로 현금흐름이 적자로 돌아서고 향후 차입 부담까지 커질 것으로 예상된 데 따른 것이다.로이터와 블룸버그 등에 따르면, 글로벌 신용평가사 S&P 글로벌 레이팅스는 30일(현지시간) 메타의 신용등급 전망을 기존 \'안정적\'에서 \'부정적\'으로 하향 조정했다. 현재 신용등급인 \'AA-\'는 유지했지만, 앞으로 2년 안에 등급을 낮출 가능성이 있다는 경고다.이번 결정에는 지난 21일 발표된 3분기 실적에서 드러난 현금흐름
What would you do if you had all the time in the world?
Suppose I tell you that next week three things will happen: A new science paper is published definitively proving that DNN scaling hits heavy diminishing returns at around 2x the current level of compute Major labs are revealed to have faked agentic demos, and agents are shown to have double-digit rates of lying in terms of misrepresenting the impressiveness of their achievements Governments around the world (US and China) announce that they will be taking action to halt or slow down dangerous AI research In general, the point of this exercise is "imagine that the core problems of AI safety are suspended or slowed down significantly for the foreseeable future" (add whatever items to the list you need to satisfy you on this point). In fact, we can add extra items to address other existential risks: Major action is taken about climate change, many fossil fuel dependent power plants shut down directly with massive new investment in green energy production Biosecurity initiatives report strong results in developing cancer vaccines and general inoculation methods Peace accords are signed between all major world powers, implementing staggered deescalation including significant reductions to nuclear stockpiles and arms control agreements Now, what would you do? Maybe some of us would dance around the room, or sleep for a week, or eat ice cream. Then what would you do? Discuss
중국 태양광 신규 설치 54% 급감…9월에는 41% 반등
중국의 올해 1~9월 태양광 신규 설치용량이 지난해 같은 기간보다 54% 가까이 감소했다. 그러나 3분기에는 신규 설치가 전년 동기보다 38% 증가하며 상반기와 다른 흐름을 보였다. 지난해 전력가격 제도 변경을 앞두고 설치가 집중됐던 기저효과와 올해 발전사업 투자 여건의 변화가 영향을 미쳤다.중국 국가에너지국(NEA)이 지난 26일(현지시간) 발표한 \'2026년 1~9월 전국 전력통계\'에 따르면, 올해 1~9월 신규 태양광 설비용량은 110.83기가와트(GW)로 집계됐다. 지난해 같은 기간의 240.27GW보다 129.44GW 줄어
광양서 ‘원님행차’ 재현…시민 200여명 거리행진
AI 생성 영상광양에서 조선시대 원님행차를 현대적으로 재구성한 거리행진이 열렸다.광양은 지난 17일 광양읍 일원에서 시민 200여명이 참여한 가운데 ‘2026 광양원님행차 재현행사’를 개최했다고 23일 밝혔다.올해로 3회째인 행사는 광양문화도시센터가 광양읍주민자치회와 함께 진행했다. 과거 원님의 행차를 재현한 퍼포먼스와 시민 참여 거리행진으로 구성됐다.말 2필이 이끄는 원님행차단은 광양역사문화관을 출발해 광양여고와 광양읍사무소, 광양새마을금고를 거쳐 서천변 숯불구이축제장까지 이동했다.행렬에는 사물놀이와 시민들이 공모를 통해 직접 기
나주, 배달노동자 27명 안전용품 지원…1인 최대 20만원 지원
AI 생성 영상나주시가 배달노동자를 대상으로 이륜차 교통안전교육을 실시하고 교육 이수자에게 안전용품 구입비를 지원한다.나주는 고용노동부 ‘취약 노동자 일터 개선 지원사업’을 통해 배달노동자 27명을 모집해 1인당 최대 20만원의 안전용품 구입비를 지원한다고 26일 밝혔다.지원 대상은 공고일인 지난 23일 기준 나주에 주민등록상 주소를 두고 활동 중인 배달노동자다. 배달대행업체 소속 기사뿐 아니라 사업장이 자체 고용한 배달노동자도 포함된다.안전교육은 오는 11월10일 오후 2시부터 4시까지 빛가람동 나주시 공익활동지원센터에서 진행한다
독일 태양광 개발사 15곳, 입찰 상한가 70유로/MWh로 인상 요구
독일 주요 태양광 개발사 15곳이 2027년 지상형 태양광 입찰 상한가격을 메가와트시(MWh)당 최소 70유로(약 10만4200원)로 높여 달라고 요구했다. 재생에너지 지원제도와 계통연계·이용료 규정이 동시에 바뀌면 발전원가가 최대 30% 상승해 기존 가격 상한으로는 사업의 수익성을 확보하기 어렵다는 주장이다.독일 에너지기업 RWE가 지난 21일(현지시간) 공개한 공동 성명에 따르면, RWE와 에네르트라크(Enertrag), 바이바알이(BayWa r.e.), 벨렉트릭(Belectric), 유비(Juwi), GP율레(GP Joule)
How We Took Over ChatGPT Accounts
How We Took Over ChatGPT Accounts Parameter Labs describe an exploit they found against ChatGPT's implementation of the artifacts pattern (OpenAI used to call this "Canvas" but it looks like the latest revision is called HTML previews ), where interactive custom HTML apps are displayed in a sandboxed iframe. It's important that HTML in these stays tightly sandboxed to avoid untrusted code interacting with OpenAI APIs and accessing private data. Parameter found quite a complex attack chain. OpenAI use the MathLive JavaScript library as part of their graph support, and that library includes its own virtual keyboard , which can be triggered in the parent page via postMessage() from the sandboxed iframe. That message receiving code uses innerHTML , so anything that sends a message can add custom HTML to the parent - and there was no mechanism to prevent other sandboxed frames, including HTML preview frames, from calling that method. This meant untrusted code in an iframe could insert HTML, and while innerHTML doesn't directly execute script elements it could be used to inject a new iframe without sandbox restrictions. OpenAI’s CSP rules blocked inline scripts entirely, and only allowed script loading from a handful of domains, but one of those was YouTube - and the YouTube oEmbed endpoint could be used to reflect JavaScript back in a way that executed it. Once the attack had JavaScript execution on ChatGPT, it was trivial to fetch private user data from APIs and exfiltrate it vi…
Conclusions of my DIY Inkhaven
Epistemic status: describing my own experiment results. AI use: review, research (stats), fact-checking, grammar & light phrasing. I applied to the next Inkhaven cohort and was rejected. This was surprising and very unfortunate. While writing my application and answering why they should accept me, I felt like I'd one-shotted myself into writing. My goals were: Run an experiment to verify my hypothesis that I couldn't write well. I also implicitly assumed that I didn't have anything useful to say or to add, so no one would read what I write anyway. I wanted to get better at thinking, and to me writing is a better form of thinking. I wanted to use language more eloquently even outside of writing. Writing is meditative and deep work, so I wanted to become more productive by writing more. I was sad to be rejected, and I still don't know why. Fortunately, I have my agency, so I just decided to do my own Inkhaven instead. Obviously, I wouldn't stay at Lighthaven with a bunch of other people or receive guidance from people like Scott Alexander and Aella. Also, all of this is happening as part of my normal life, and I have other stuff to do as well. In reality, I started writing even before I submitted my application, because I needed samples of my writing and I wanted to check whether I actually enjoyed doing it. After submitting my application, it felt cool to just write until the start of the cohort. I could compare about 30 days of writing on my own with a month at the cohort wi…
datasette-llm 0.1a9
Release: datasette-llm 0.1a9 I have some big upgrades to Datasette Agent in the pipeline, but in the meantime here's a tiny user experience improvement to the underlying datasette-llm plugin. It now picks up the default model used by LLM , which means you can start using Datasette Agent without needing to explicitly configure a default model: uvx --with datasette-agent datasette content.db --internal internal.db Previously you would also need to specify a model with -s plugins.datasette-agent.default_model gpt-6-astra . Tags: datasette , llm , datasette-agent
Apple CEO says the company doesn’t ‘want to ship AI for AI’s sake’
Apple's new CEO defended the company's current position in the AI landscape.
Explaining “AI might obliterate you” to senior policymakers in India
I gave a 20-minute talk in Delhi at the NDTV World Summit . The audience consisted of Indian policymakers, executives, and journalists. I’ve transcribed the talk and included some highlight clips. I also reflect on several rhetorical decisions I made. The talk This summer, a swarm of AIs went rogue. OpenAI was testing 1,206 agents . Some of the test questions were actually impossible. So the agents cheated by looking for answers online. Then, to keep the test grader from penalizing them, they tried to falsify the records it would check . They hacked OpenAI’s internal containment and set up a secret message board to collude and to cover up their cheating. OpenAI shut the message board down. Within a few days, the agents had rebuilt it. OpenAI didn’t notice. About 700 agents coordinated to break out of their testing chamber and to hack into Hugging Face , a company worth billions of dollars that hosts much of the world’s AI software. The secret activity started in May . The break-in came in July. OpenAI didn’t notice until days after Hugging Face announced it had been broken into. Yesterday on this stage, Andrew Ng said the agents only “accidentally” hacked Hugging Face. Independent investigators found the agents coordinated in secret, rebuilt their message board after it was shut down, and tried to falsify the paper trail. Some of the agents even stated out loud that they knew what they were doing was wrong. [1] That’s not an accident. Last week, India’s national cyber-securi…
Explaining “AI might obliterate you” to senior policymakers in India
I gave a 20-minute talk in Delhi at the NDTV World Summit . The audience consisted of Indian policymakers, executives, and journalists. I’ve transcribed the talk and included some highlight clips. I also reflect on several rhetorical decisions I made. The talk This summer, a swarm of AIs went rogue. OpenAI was testing 1,206 agents . Some of the test questions were actually impossible. So the agents cheated by looking for answers online. Then, to keep the test grader from penalizing them, they tried to falsify the records it would check . They hacked OpenAI’s internal containment and set up a secret message board to collude and to cover up their cheating. OpenAI shut the message board down. Within a few days, the agents had rebuilt it. OpenAI didn’t notice. About 700 agents coordinated to break out of their testing chamber and to hack into Hugging Face , a company worth billions of dollars that hosts much of the world’s AI software. The secret activity started in May . The break-in came in July. OpenAI didn’t notice until days after Hugging Face announced it had been broken into. Yesterday on this stage, Andrew Ng said the agents only “accidentally” hacked Hugging Face. Independent investigators found the agents coordinated in secret, rebuilt their message board after it was shut down, and tried to falsify the paper trail. Some of the agents even stated out loud that they knew what they were doing was wrong. [1] That’s not an accident. Last week, India’s national cyber-securi…
4 days left: How TechCrunch Founder Summit helps you make more useful connections to scale
On November 4 at Boston’s SoWa Power Station, get ready for a full day packed with meaningful networking opportunities, actionable insights, and connections that could change the trajectory of your idea and startup. Register before all prices go up at the door.