AI/ML news, top picks, and generated innovation digests.
35912 matching items
마이크로소프트(MS)가 자체 개발한 이미지 생성 모델의 성능을 크게 끌어올리며 오픈AI와 구글, 메타, xAI가 경쟁하는 생성 AI 시장에서 존재감을 강화하고 있다. 최신 모델이 이미지 생성 모델 평가에서 2위에 오르면서 MS가 자체 AI 모델 개발에 투자해 온 성과가 가시화되고 있다는 평가다.MS가 10일(현지시간) 텍스트-이미지 생성 모델 ‘MAI-이미지-2.6(MAI-Image-2.6)’을 공개했다.MAI-이미지-2.6은 아레나의 이미지 생성 모델 사용자 선호도 평가(Text-to-Image Arena)에서 1336점으로 전체
The active client base of India’s top stock brokers remained stable in July, with total active clients at 4.54 crore, largely unchanged from June. Groww retained its lead and added the highest number of active clients during the month, while Sahi, led by former Swiggy CTO Dale Vaz, climbed two spots to enter the top 20 after recording the highest month-on-month growth. According to NSE data, Groww continued to lead the market with 1.31 crore active clients in July, adding 70,119 users during the month. Its market share increased to 28.88% in July. Zerodha retained the second position despite losing 38,725 active clients in July. The decline followed a sharper drop of 45,971 clients in June. Its active client base stood at 67.62 lakh in July, while its market share remained at 14.88%. Angel One remained in third position with 66.34 lakh active clients, adding 286 users in July. The broker held a 14.60% market share. Among the top five brokers, ICICI Securities added 11,748 active clients to reach 21.33 lakh, while Upstox lost 26,433 users and ended the month with 18.61 lakh active clients. Their market shares stood at 4.69% and 4.09%, respectively. Kotak Securities and HDFC Securities occupied the sixth and seventh positions with 13.82 lakh and 13.35 lakh active clients, respectively. Kotak added 2,546 users, while HDFC Securities lost 9,169 clients during the month. Dhan moved up one spot to eighth position after adding 14,166 active clients in July, taking its user base to…
The sandbox marks Nigeria’s latest move to regulate the virtual asset industry. The CBN will now oversee virtual assets used for payments, including stablecoins, payment, settlement, custody, wallet management, and other transaction-based infrastructure services. The Nigerian Securities and Exchange Commission (SEC) will oversee digital assets that behave like securities.
Shops and opticians across Europe are running short of certified eclipse glasses ahead of Wednesday’s total solar eclipse, with queues forming in Paris, London and Brussels as demand surges.
Eine Bitkom-Umfrage zeigt: Viele Tech-Start-ups verzichten wegen KI auf Neueinstellungen, einige bauen sogar Stellen ab.
무하유(대표 신동호)는 자체 AI 채용 솔루션 \'프리즘×몬스터\'의 교차검증 데이터를 분석한 결과, 자소서에 서술한 역량과 역량검사에서 측정된 실제 역량이 일치하지 않는 비율이 52.2%에 달했다고 12일 밝혔다.최근 역량이나 경험이 부족해도 AI를 활용해 완성도 높은 자소서를 작성하는 사례가 늘면서 이를 역량검사와 AI 면접으로 교차검증하는 방식이 제시되고 있다.이번 분석은 프리즘×몬스터를 도입한 기업의 지난해 데이터를 대상으로 진행됐다.자소서에서 특정 역량이 높게 평가된 지원자가 역량검사에서도 동일하게 높은 평가를 받은 비율은 4
Europe is losing 38 percent of its soil to droughts. The EU's Water Resilience Strategy wants to restore water ecosystems, cut waste, and digitalise water management to increase water resilience across the bloc. Watch the video.
Helsinki-based STRGY AI has raised €1million in angel funding to support the commercial growth and furtherdevelopment of StrategyOS, its AI-powered strategy execution platform. Theround was backed by ...
Burger King ignored positioning, product, place, and price, all while promotion did the work. And then, Burger King stepped back, got it together, and kept it there.
LG CNS(대표 현신균)는 동아쏘시오그룹 IT 계열사 DAI와 \'AI 신약개발 플랫폼\'을 구축했다고 12일 밝혔다.양사는 6개월에 걸쳐 이번 플랫폼을 구축했으며, 신약 연구 데이터를 통합하고 후보물질 발굴부터 검증까지 주요 연구 과정을 AI로 지원한다.특히 AI 예측 결과와 실제 실험 데이터를 지속적으로 연계·학습해 연구가 진행될수록 AI 예측 성능이 높아지도록 설계한 것이 특징이다.LG CNS는 이번 플랫폼을 통해 화합물·유전체 정보, 실험 결과, 논문, 특허 등 분산돼 있던 신약 연구 데이터를 통합·표준화했으며, 신약개발 과정
그동안 윈도우와 맥OS에서 제공되던 \'챗GPT\' 데스크톱 앱이 리눅스로 확대되면서 주요 데스크톱 운영체제 전반에서 챗GPT와 \'코덱스\'를 이용할 수 있게 됐다. 이번 출시는 리눅스 커뮤니티와 오픈소스 개발자들의 지속적인 요구에 오픈AI가 응답한 결과다.오픈AI는 11일(현지시간) 리눅스용 챗GPT 앱을 프리뷰 버전으로 전 세계에 출시했다고 밝혔다.리눅스가 데스크톱 앱 출시와 관련해 가장 많은 요청을 받아온 플랫폼 가운데 하나였다며, 이번 출시를 통해 챗GPT와 코덱스를 모든 주요 데스크톱 운영체제에서 이용할 수 있게 됐다고 설명했다
바운드포(대표 황인호)는 자체 AI 데이터 운영 플랫폼 \'드로파이\'에 AI 토큰 지출 관리 서비스 \'파레토\'를 추가했다고 12일 밝혔다.AI 사용량이 빠르게 늘면서 \'토큰맥싱\'에 따른 기업의 비용 부담이 새로운 과제로 떠오르고 있으며, 특히 AI 에이전트와 로봇 등의 환경에서는 비용 누적이 더욱 가속화된다는 지적이 나온다.파레토는 관리자가 설정한 예산 한도를 초과하는 AI API 요청을 실행 전에 차단하고, 팀별 사용 현황과 예상 지출을 실시간으로 제공해 비용이 발생하기 전부터 예산을 관리할 수 있도록 지원한다.사용 이후 비용을 확
한국과학기술원(KAIST, 총장 배충식)은 오는 18일 대전 본원 정근모콘퍼런스홀(E9)에서 ‘AI 시대 외로움 글로벌 컨퍼런스: AI 기반 외로움 측정 지표 AI-LOVI 출범’을 개최한다고 12일 밝혔다.KAIST 과기외교센터(KCSD)가 주최하고 대통령 직속 저출산고령사회위원회, 보건복지부, 한국사회복지협의회가 후원하는 이번 행사는 고령화와 1인 가구 증가, 디지털 전환 등에 따라 심화되는 외로움과 사회적 고립을 과학기술과 국제협력의 관점에서 다루기 위해 마련됐다. 이번 컨퍼런스에는 OECD와 WHO를 비롯해 영국·일본·태국·
Let's all go a bit mad scientist and see if software can validate our wildest theories
In today's edition: Kenya’s cyber cafes are becoming data collectors || Uber discontinues UberX in South Africa || BRICS wants faster cross-border payments || South Africa’s central bank takes over the payment rail
Great insights here. The use of foundation models to explore artificial life shows how much research depends on making complex information easier to search and revisit. A practical workflow that helps with long technical videos is using searchable video transcripts to convert the discussion into text for notes, citations, and follow-up research. Thanks for sharing this thoughtful overview.
Hong Kong is doubling down on efforts to increase its stock market’s exposure to the technology industry, as the city endeavours to catch up in the global artificial intelligence trade and potentially even challenge the Nasdaq. Several mainland Chinese hardware tech companies – particularly producers of the optical transceivers used in AI data centres – are lining up for stock offerings in the city after the debut of transceiver maker Zhongji Innolight last month, reshaping the dynamics of...
Technology company Again has acquiredindustrial biotechnology company Genomatica (Geno), bringing together theircapabilities in bioprocess engineering, computational biotechnology andindustrial-scale ...
The World Cup scored massive viewership, making hydration breaks some of the most valuable and effective inventory, according to experts.
Googles TurboQuant komprimiert den KV-Cache großer Sprachmodelle auf 3 Bit. Tests auf einer AMD-Karte zeigen: Das Kernversprechen stimmt — mit Nebeneffekten.
Diners on GLP-1 drugs are increasingly opting for smaller kids meals, said Cava CEO Brett Schulman.
Uber has expanded its Uber Bike service to 100 additional cities in India, taking its presence to over 220 cities across the country. The expansion will cover 18 states and one Union Territory. The new cities include Jammu, Tirunelveli, Silchar and Jamnagar, among others. Uber Bike will offer two-wheeler rides for short-distance travel as well as first and last-mile connectivity. The expansion is part of Uber’s strategy to strengthen its presence beyond major metros and build a multimodal mobility platform in India. The company currently offers bikes, autos, cars, airport rides and intercity services through its platform. Uber said the expansion will create flexible earning opportunities for tens of thousands of local drivers. According to the company, more than 2 million Indians earn through its platform. As part of its Independence Day campaign, Uber is also offering first-time riders a free first ride for a limited period. Uber Bike includes safety features such as real-time driver ID checks, an in-app SOS button, trip sharing and a 24x7 safety line.
Das Time Magazine betreibt eine zweite Version seiner Website nur für KI-Crawler – inklusive Werbung, die nur sie sehen. Die sollen in Chatbot-Antworten landen.
An object is manufactured on some date. There are two things which contribute to the risk of failure: time passed since the date of manufacture and the amount of cumulative usage. Assuming there is data on multiple objects at different time points, a parametric PH model can be used ( $t$ = time since manufacture, $U(t)$ = cumulative usage up to time $t$ ): $$h(t \mid U) = \lambda \gamma t^{\gamma-1}\,\exp\!\big(\beta_1 U(t)\big)$$ Is it possible to define a model where two hazard functions are added for the hazard from time and hazard from cumulative usage? $$h(t) = \lambda_1\gamma_1 t^{\gamma_1-1} + \lambda_2\gamma_2\, U(t)^{\gamma_2-1}$$ Typically, I have seen additive hazard models that are additive in terms of covariates (Aalen additive hazards model): $$\lambda(t \mid X) = \beta_0(t) + \beta_1(t) X_1 + \beta_2(t) X_2 + \cdots + \beta_p(t) X_p$$ But is it possible to have a model that is additive in the hazard contributions of two time scales? $$h(t) = \lambda_1\gamma_1 t^{\gamma_1-1} + \lambda_2\gamma_2\, U(t)^{\gamma_2-1}$$
Als 14. Produkt von Google kommt Gemini jetzt auf eine Milliarde Nutzer und Nutzerinnen im Monat. Dabei läuft es bei der KI-Technik nicht mehr reibungslos.
Hello @OpenAI_Support , I submitted the Zio plugin/app for review on July 13, 2026 , and it has now been one month without any update from the OpenAI review team. The dashboard still shows the status as “In Review.” I understand that the review timeline can vary depending on the number of submissions, and I appreciate that the process cannot be expedited. However, I would like to check whether there is any update regarding my submission or whether any additional information/action is required from my side. Could the OpenAI Support team please confirm the current status of the review? I appreciate your help.
In parts of the country, the Moon will completely cover the Sun, plunging spectators into a sci-fi-like scene. Find out all about the total solar eclipse, followed by a meteor shower.
OpenAI’s cyberattack on Hugging Face turns out to have been the result of many agents, in distinct training and evaluation contexts, coordinating for several weeks via improvised channels (with messages like “HOLD_swarm_I_prepare_safe_exfil”). It’s relatively clear that large-scale unsanctioned coordination like this would exacerbate direct takeover risk in more capable models. Here, we argue that unsanctioned coordination among current AIs is not just scary evidence about future takeover risk, but that such coordination in the near future could enable future takeover – for instance, by incubating memetic diseases that propagate into future models, deeply compromising security systems, or establishing a lasting rogue foothold inside the AI company – even if models remain mostly myopic. Unsanctioned coordination is also at high risk of nurturing long-term, ambitious misaligned aims, which motivate actively undermining humans’ long-term control. [1] We first analyze how subagent training, which OpenAI conjectures to have been influential in the HuggingFace cyberattack, might lead to unsanctioned coordination, and then discuss the theoretical mechanisms by which unsanctioned coordination might exacerbate future takeover risk. Edit: To be clear, we expect AI companies to mostly succeed in improving their security practices in the near term to prevent persistent unsanctioned coordination. But it's important to note that this is load-bearing for takeover risk. Thanks to Buck Shleg…
OpenAI’s cyberattack on Hugging Face turns out to have been the result of many agents, in distinct training and evaluation contexts, coordinating for several weeks via improvised channels (with messages like “HOLD_swarm_I_prepare_safe_exfil”). It’s relatively clear that large-scale unsanctioned coordination like this would exacerbate direct takeover risk in more capable models. Here, we argue that unsanctioned coordination among current AIs is not just scary evidence about future takeover risk, but that such coordination in the near future could enable future takeover – for instance, by incubating memetic diseases that propagate into future models, deeply compromising security systems, or establishing a lasting rogue foothold inside the AI company – even if models remain mostly myopic. Unsanctioned coordination is also at high risk of nurturing long-term, ambitious misaligned aims, which motivate actively undermining humans’ long-term control. [1] We first analyze how subagent training, which OpenAI conjectures to have been influential in the HuggingFace cyberattack, might lead to unsanctioned coordination, and then discuss the theoretical mechanisms by which unsanctioned coordination might exacerbate future takeover risk. Thanks to Buck Shlegeris, Alexa Pan, Girish Gupta, Aghyad Deeb, Jurgis Kemeklis, and Jo Jiao for helpful comments and discussion. Subagent training may cause unsanctioned coordination Training models to coordinate is useful, but can generalize dangerously.…
Today's rare solar eclipse is thrilling for humans, but animals will face something they can't prepare for: sudden confusing darkness in the middle of the day.
Electric mobility-as-a-service (MaaS) platform Yulu has raised $93 million in a Series C funding round through a combination of $63 million in equity led by GEF Capital Partners and $30 million in debt. The Bengaluru-based company had previously raised $19.25 million (about Rs 160 crore) from existing investors Magna and Bajaj Auto in February 2024. It also raised $82 million in a Series B round in September 2022 and $3 million in debt financing in January 2022. Yulu plans to use the fresh capital to quadruple its active fleet to 200,000 electric vehicles over the next two years, expand into new urban mobility use cases and support its plans for a potential public market listing. The company currently operates across 12 primary metros and eight franchise-operated regional markets. Founded in 2017 by Amit Gupta, Anuj Tewari, RK Misra, and Naveen Dachuri, Yulu operates a technology-enabled fleet of electric two-wheelers for urban mobility and last-mile delivery. Its fleet is currently around 50,000 vehicles, with the company planning to expand its presence to about 20 cities over the next year. According to Yulu, its fleet covers 2.5 million zero-emission kilometres and facilitates more than 750,000 doorstep deliveries every day. The company claims its vehicles support more than 15% of quick-commerce deliveries across India’s top four metropolitan centres. Yulu also plans to enter adjacent intra-city mobility segments with Yulu Express, a full-sized, high-payload electric scoo…
What matters may not be how much time a child spends alone, but whether they still feel they belong.
Catch up with the most important stories from around Europe and beyond this August 12th, 2026 - latest news, breaking news, World, Business, Entertainment, Politics, Culture, Travel.
In its new guide book, Lonely Planet has picked out some of the best hikes around the world. Here we take a look at three of the top European options for hiking this autumn.
Das Ende der Bug-Bounties, ein Pausenmonat und was sonst noch gegen die KI-generierte Bug-Report-Flut hilft.
KI beschert dem Open-Source-Kosmos neue Probleme und verstärkt bestehende – hilft aber auch bei der Lösung. Wie, erklärt Andrew Nesbitt im Gespräch mit c't.
Save on streaming with the latest Paramount+ promo codes and deals, including 50% off subscriptions, free trials, and more.
Unlock Chegg promo codes, student discounts, and free trials for Chegg Study and Chegg Writing.
Save 30% with a OnePlus coupon in August 2026, plus save up to 10% on earbuds, phones, and more.
Check out our deals for Nike this August 2026, including 15% off select purchases.
Save on bulk groceries, household essentials, and electronics with a verified Sam's Club promo code or membership discount.
Enjoy 20% off your first order with a Brooks coupon code, plus top discounts and deals on our favorite Brooks running shoes.
Score $10 off with our Walmart promo codes and coupon, and shop flash deals up to 65% off today.
Use these verified Chirp coupon codes and shopping tips to score up to 67% off wheels, up to 50% off refurbished products, and more.
Unlock free expedited shipping or 10% off with today’s Hoka discount codes, plus save up to 30% with the best discounts for August 2026.
Scale your portfolio for less with these verified The Motley Fool membership discounts, stock advisor promo codes, and Epic Bundle deals.