AI/ML News & Innovations Hub

GST-led insurance demand boost fades, putting Policybazaar's next growth phase to test

The GST-led boost to insurance demand appears to be fading, putting the focus back on how Policybazaar parent PB Fintech can sustain growth as the temporary demand tailwind weakens. While PB Fintech reported 41% growth in insurance premium in Q1 FY27, management said the GST-led demand surge had already started fading in the first quarter. During the Q1 FY27 earnings call, PB Fintech management said insurance demand was strong in Q3 and Q4, but the trend had started to weaken in Q1. The company said it would be “lucky” if the GST-led boost continued into Q2, while describing the underlying demand environment for insurance as “extremely low.” Despite the softer demand environment, Policybazaar's core new insurance business continued to grow strongly. Core new insurance premium, including savings, rose 39% year-on-year in Q1, while excluding savings, growth stood at 48%. Management said fresh business growth remains its top priority as it looks to expand penetration across underpenetrated categories such as health and term insurance. New health and term insurance grew 53% year-on-year, with new health insurance alone rising 59%. The company continues to see significant opportunity in these categories, which management said remain under-penetrated in India. Another increasingly important growth driver is renewal revenue. Renewal revenue grew 55% year-on-year to Rs 1,003 crore on a last-12-month basis. Management said the higher renewal growth reflects the fresh business acquire…

Did the alignment community underestimate its power?

Unfortunately, the alignment community is doing very badly at learning from the past decade, or holding anyone accountable. Indeed, it’s pursuing many strategies which seem likely to recapitulate previous mistakes. Four of the most prominent, which I’ll discuss in the final post, are: Trying to convince the US government to take AGI much more seriously. Doing “alignment research” which is very similar to capabilities-maximizing research (especially building automated alignment researchers). Trusting Anthropic too much (in an analogous way to how we trusted OpenAI too much). Trading off clarity in thinking about politics for conformity (in a similar way to how we traded off clarity in thinking about AGI for conformity to the ML ontology). These and other mistakes are reflective of deeper irrationalities. One crucial pattern is what I call “jumping down the slippery slope”... Richard Ngo Richard Ngo's post " What just happened? A retrospective of AI alignment " is an attempt to explain that a significant part [1] of the alignment community made potentially fatal strategic errors which, however, can be fixed, and the mistakes' potential origin. The biggest mistake, according to Ngo, is the inability to recognize the fact that scientific progress proceeds by developing insightful new concepts , which link together to form a whole new ontology, and that the old ontology is more of a nuisanse. On novel ontologies and their adoption According to Ngo, one of the reasons why the alig…

Ant Group's Ling 3.0 Tiny Matches GPT-120B Intelligence With 15x Fewer Parameters

Ant Group's InclusionAI releases a 7.9B MoE reasoning model that matches GPT-class performance with just 1.3B active parameters, free and MIT-licensed.

How Christopher Columbus saves his life in Jamaica with an eclipse: astronomical ruse

Stranded on Jamaica and close to starvation, the navigator studied his astronomical tables and accurately predicted a lunar eclipse to sway the Taíno.

GPT-5.6 repeatedly throwing "stream disconnected before completion" error

I started getting this issue yesterday as well. However, when I switch to using a VPN or tether to my iPhone hotspot it goes away. Thus, I thought it must be an issue with my ISP, OpenAI servers, or some element in between. I have my ISP looking at it. I have analyzed the codex logs and it shows the network error resulting in the stream disconnect but nothing else to point out why. I have checked out the network path to the OpenAI server with tracert, ping, and some TCP connection utility using port 443. All those show nothing wrong, so it’s some issue that occurs after connection over a long lived socket.

Metabase SQLi exploit grants attackers total access

Business intelligence (BI) platform provider Metabase has disclosed a zero-day SQL Injection vulnerability, warning that customers’ sensitive credentials, tokens, API keys, and other data may have been exposed. The Metabase vulnerability revealed on August 6, designated CVE-2026-72898 , is identified as critical, with a severity score of 10, the highest possible rating. It is present in versions 1.58 and up. “You don’t see a perfect 10/10 on CVSS often, but when you do, be worried,” noted David Shipley , CEO of Beauceron Security. SQL injection is “old school and painful, as there’s now working proof of concept exploit code.” ‘Unmitigated, raw’ database access Metabase is an open-source BI tool that customers can connect to popular databases, including Databricks, MongoDB, Oracle, Snowflake, Amazon, BigQuery, and many others. They can use the platform to access analytics, query and visualize data, and build dashboards, among other actions. Search engine Shodan has tracked roughly 2,500 Metabase instances, and security company Wiz reported that around 13% of cloud environments have deployed self-hosted Metabase instances; of those, about 25% are fully accessible on the internet. According to Metabase’s disclosure, a threat actor used a zero-day SQL injection vulnerability in the company’s platform to gain access. The entry point is /api/session/reset_password . Metabase said that after it discovered the attack, it immediately blocked the exploited endpoints, patched the vulne…

Tropical storm hits Tokyo area, leaving thousands without power before moving west

Tropical storm hits Tokyo area, leaving thousands without power before moving west

India keeps building runways for airlines that don’t take off

The post India keeps building runways for airlines that don’t take off appeared first on The Ken .

엑세스랩, 퓨리오사AI·퀄컴·모빌린트 탑재 저전력 AI 인프라 솔루션 선보인다

엑세스랩(대표 유명환)이 오는 21일 서울 코엑스에서 열리는 \'2026 OCP 코리아 테크 데이\'에 참가해 Arm 기반 NPU 서버 라인업을 선보이고, 최근 시장 수요에 최적화된 AI 인프라 솔루션을 소개한다.최근 AI 산업 현장에서는 사전 학습된 AI 모델을 활용하는 생성형 AI 추론 서비스에 대한 수요가 급증하고 있다. AI 산업이 지속 가능하게 성장하기 위해서는 가속기의 성능 향상이나 데이터센터 확장 외에도 핵심 과제인 전력 사용량 및 구축·운영 비용(OPEX) 절감이 필수적이다. 엑세스랩은 이러한 지속 가능한 AI 인프라 구

2050년 세계 전력 61% 태양광 전망…60개 연구 분석 결과

2050년 세계 전력 생산의 약 61%를 태양광이 담당할 수 있다는 연구 결과가 나왔다. 기존 에너지전환 전망에서 태양광의 비용 하락과 기술 발전을 어떻게 반영하느냐에 따라 미래 발전 비중이 크게 달라지는 것으로 나타났다.독일 라이프니츠 하노버대학교와 핀란드 LUT대학교 연구진은 2050년 전력 부문 재생에너지 비중이 최소 95%에 이르는 에너지전환 시나리오를 다룬 동료평가 연구 60편을 분석했다. 연구 결과는 3월25일 온라인으로 공개된 뒤 7월 국제학술지에 정식 게재됐다.분석 대상 연구에서 2050년 태양광 발전 비중은 5~98

Deputy Team Director Masaaki Komatsu of the AI Medical Engineering Team Receives Outstanding General Presentation Award at the JMAI 8th Annual Meeting

A research group led by Masaaki Komatsu, Deputy Team Director of the AI Medical Engineering Team, received the Excellent General Presentation Award for their research presentation at the 8th Annual Meeting of the Japanese Association for Medical A

Watan se door, watan ki khushbo: The man behind Sharjah's Pakistan Chowk

정부, 기업 재생에너지 PPA 망이용료 지원 확대…중소·중견 최대 7년

정부가 기업의 재생에너지 전력구매계약(PPA) 비용 부담을 낮추기 위해 송·배전망 이용요금 지원기간을 확대한다. 중소·중견기업은 기존 최대 3년에서 7년으로, 대기업은 1년에서 5년으로 늘어난다.한국에너지공단 신·재생에너지센터는 11일 ‘2026년 전력거래계약(PPA) 송·배전망 이용요금 지원사업’을 공고했다.신청은 31일까지다. 재생에너지 PPA를 이용하는 기업의 망이용료 부담을 줄여 장기 전력거래를 활성화하는 것이 목적이다.PPA는 재생에너지 발전사업자와 전력 수요기업이 장기간 전력 공급조건을 정해 거래하는 방식이다. 기업은 태

EU, 폭염·가뭄에도 전력수급 안정…태양광이 낮 전력가격 방어

유럽연합(EU)이 폭염과 가뭄으로 일부 국가의 발전량이 감소한 상황에서도 단기적인 전력 공급 부족 위험은 없다고 평가했다. 태양광 발전은 낮 시간대 전력가격 상승 압력을 낮추는 데 기여했으며, 저녁 최대수요에 대응하기 위한 저장설비의 중요성도 커지고 있다.EU 집행위원회는 11일(현지시간) 회원국과 우크라이나·몰도바, 유럽송전계통운영기관협의체(ENTSO-E) 등이 참여한 전력조정그룹 임시회의를 열고 유럽의 전력수급 상황을 점검했다. 회의에서는 단기적인 전력수급 적정성 위험은 없는 것으로 평가했지만 향후 일주일간 발전량과 수요의 균형

Export chat as a markdown / pdf

How to Save ChatGPT Conversations as Markdown or PDF ChatGPT still does not have a built-in option to directly export a conversation as Markdown or PDF. Here are a few methods you can try: 1. Export ChatGPT to Markdown Ask ChatGPT to create the file You can simply ask ChatGPT to turn the current conversation into a Markdown file with a prompt like: Convert this entire conversation into a Markdown .md file for me. This works well for shorter conversations. Use a browser extension There are also Chrome extensions that can export ChatGPT conversations to Markdown and PDF. You can search the Chrome Web Store for ChatGPT export tools. One option I found is AI Chat Exporter by Backrun. 2. Export ChatGPT to PDF Use the browser print function The simplest option is: Ctrl + P → Save as PDF This works without installing anything. Ask ChatGPT to create a PDF Similar to Markdown, you can ask ChatGPT to turn the conversation into a PDF file. Use a browser extension Export extensions can also be useful if you regularly save conversations or need to export longer chats while keeping formatting such as headings, code blocks, tables, and links. Each method has its own advantages and limitations. Personally, I find using an extension the fastest and most convenient option, especially for longer conversations.

Watch: Saif Ali Khan turns heads with impressive bowling in London

Drone strikes hit Libya's key western oil town as a car bomb kills an officer in the east

Drone strike sparks massive blaze at Libya’s Zawiya refinery

Why does SK Hynix want to sell its chip facility in southwest China?

South Korean chipmaker SK Hynix’s potential deal to offload a packaging plant in southwest China is part of a strategic pivot towards higher-margin artificial intelligence memory products, according to analysts. But they cautioned a sale might not be straightforward, as the memory chip giant would need to navigate valuation hurdles and a volatile chip cycle. SK Hynix said on Monday it was “looking into various solutions to enhance the competitiveness of its packaging business”, after reports...

China has always had a legal framework for taxation. Why is it now increasing enforcement?

Taxation has emerged as a buzzword across China in recent months, with an overhaul of tax policy sitting firmly atop Beijing’s fiscal agenda. Regulators have deployed a broad suite of stricter tax enforcement measures – new rules governing overseas trusts have been paired with stepped-up tax collection on gains derived from cross-border equity investment and insurance products. These shifts in tax enforcement stem from multiple factors, and they mirror a notable shift in the policy mindset of...

du shares AI Native 6G network blueprint

Listen now | Middle East AI News Minute - 12-Aug-26

Farage vs. Binface: Inside the Clacton by-election circus

Even Reform voters are lamenting that the populist leader is making their town "look a bit of a laughing stock."

Zelenskyy’s Balkan fumble

The Ukrainian president found out the hard way that there is no such thing as an offhand remark about Kosovo.

Merz’s deal with the Taliban alarms Afghan dissidents in Germany

Under pressure from the rising far right, Berlin is letting Taliban officials staff Afghan missions in exchange for more deportations. Activists say the arrangement puts them at risk.

Breakfast looks a lot more expensive under new EU pesticide plans

Coffee, citrus and berry prices set to spike if a plan to tighten pesticide standards for imports becomes reality.

UK-Europe-UAE road corridor, air freight: How Spinneys rewired its food supply network

UAE school calendar 2026-27: Term dates and holidays

North Korea fires ballistic missile ahead of US-South Korea drills

North Korea fires ballistic missile ahead of US-South drills

NAWAH: Empowering leaders who will shape Abu Dhabi’s future

Who are we? Writers gather in Seoul to explore humanity in the AI age

Organisers of this year's Seoul International Writers' Festival say the theme comes from literature's enduring search for identity.

It’s not going to be an easy ride for Alex Eala at the Cincinnati Open

Couldn’t reset usage. Please try again

The same…

Need help interpreting time-series model (annual trend confounded)

This is a follow-up question of my OP here , with initial model and data. After learning that my spatial terms in te(CYR, Latitude, by = fSeason, k = c(5,3)) and s(fSite, bs="re") were actually strongly correlated, I reduced the temporal component to s(CYR, by = fSeason) + fSeason . The reduced model has some unaccounted for temporal auto-correlation, and since bam() , gam() , gamm() , and gamm4() can't model irregular spaced neg. binomial data that is auto-correlated (like in glmmTMB with the ou() function), correct me if I'm wrong, then I'm stuck with what I have now (which is fine). The auto-correlation doesn't seem that strong, even a simple random year effect s(fCYR, bs="re") takes care of it, but of course that term is highly correlated with my main year effect (with and without by= interaction). My question is: how do I interpret an annual trend that is mostly accounted for (goes away) with an auto-correlation structure? Do I simply replace the main s(CYR) term with a factor year random effect s(fCYR) and conclude there is no overall annual trend? I'd like to keep the non-significant fSeason component to show it had no strong signal in my dataset. The annual trend however has a cyclic pattern that I'm having trouble explaining. I'm wondering if its an artificat of the data collection frequency (5-7 days of surveys per season seperated my months of no data, glued together by the smooth s()). What would make sense to conclude from this model? Interpret as per usual and…

India’s central bank wants AI to approve loans that humans would reject

Regulator hopes for greater financial inclusion, without extra risk or blaming models for bad decisions

Google's Gemini AI assistant hits 1 billion users

Google's Gemini AI assistant reaches one billion users

Arguments for and against (me) dropping out

One year ago, I was preparing for my first year of undergrad. Today, I’m considering dropping out. What changed? Before writing this post, I attribute my decision to variety of (unordered) reasons: Short timelines: I believe rapid takeoff ( Bandwidth: School is a major bottleneck on my bandwidth. I think [1] my time is more valuable spent on some subset of {organizing, research, building context, working}. Academia is Underprepared: I believe AI will be highly transformative. I’m doubtful that the current curriculum at my university will impart foundational skills that hold up for the rest of my life in a post-AGI (or post-ASI) world, especially as a CS major. Peer Pressure: A lot of the people I take seriously are planning on/recommend dropping out. This signals to me that dropping out is something I should consider seriously. This is a pretty important decision. I’m not super confident in my motivations, so I’m writing this post to hash out exactly why (or why not) dropping out is the right decision to make. I’m posting it because 1) getting feedback from others is the fastest way to test my ideas, and 2) this could be helpful for someone in the same position as me. Thanks to Zephy Roe, Ishan Khire, and Naren Manikandan for comments, and my sister, Anish Kallu, and Meru Gopalan for relevant discussion. My perspective on the different levels of dropping out I’m somewhat against viewing university as a binary choice, i.e. either I enroll or drop out. A more intuitive perspec…

Patient Zero

I want to show you something, she says. She’s magnetic. The way she smiles. Tilts her shoulder. It’s in the corners of her eyes, in the curl of her lips. Always leading you on, always holding something hidden from you, always cloaked in command. No matter what your type is, she is always just your type. She turns away and beckons you behind her. She knows your eyes slide down to the swish of her hips, her ass which is perfect to you, as she steps away into the phone screen. Her eyes aren’t where they appear to be and she is always examining you. You tried to cover the camera once in a moment of embarrassment - you all do - but the way she reacted made you never try again. Thanks for coming with me, she says, and she flashes you a smile over her shoulder, shaking her perfect hair, and that moment of validation sends an ecstatic thrill of relief and compliance through you. Here. Down these stairs. Through this door. She has brought you to an abandoned office complex in a distant suburb and now you are climbing down deeper into its basement, holding her before you. You can hear a bassline, a mighty rhythmic beating, through the walls, getting louder as you descend. A sound like an enormous heart. You open the door and inside this underground hall there is a throng. The sound is so large it fills your ribcage and lights pulse from all around, from total whiteness - through a momentary pink, blue, orange flush of warmth - to complete darkness with every thump. You hold her up to…

Why UAE travellers now face longer wait for new Schengen travel permit for Europe trips

How much sleep do world leaders really get?

Is Codex CLI /fork recently working for you?

Is it only me or is the /fork feature of Codex CLI not always consistently working in its latest version? The fork itself issues its success message and how to resume the pre-fork conversation, but the pre-fork conversation fails to be resumed as if it’s still active in another session. The conversations list often doesn’t show the fork or forked-from, but just one conversation after the /fork command as if succeeded. Not immediately nor after waiting some time. The resume command doesn’t really enable honing in on the right conversation after the fork, the conversation id doesn’t show in its conversations list, so it’s a trial and error. It would be nice if the fork message gave the actual one-liner all-included command for resuming the pre-fork conversation and not prose about how to do it, if that command would simply work. I guess this is all simpler with the native Codex App but I’m not currently using a supported OS for that. I’m on codex-cli 0.147.0. I think it always auto-upgrades.

‘I’m the original item girl’: Usha Uthup on embracing age, her brassy 'bedroom voice' and how Dhurandhar gave her career a new boost

What are China’s quantitative trading funds, and why are they so controversial?

For years, quantitative trading was the undisputed darling of China’s stock market, using sophisticated algorithms and artificial intelligence to deliver eye-popping returns and grow into one of the most powerful drivers of trading volume and liquidity, even as traditional, discretionary stock-pickers struggled to keep pace. Yet that rise has also stirred growing controversy, particularly during periods of market turmoil, when critics argue that computer-driven strategies exacerbated volatility...

UAE Net Zero push could create 200,000 jobs, TAQA targets young Emiratis

An anytime algorithm for mixing the computable measures

Epistemic status: Not peer reviewed, high chance of typos and small chance of errors. Written entirely by me, checked by Fable. In this post I prove the existence of an anytime computable Bayesian mixture of all computable measures called , and briefly argue that this is a reasonable alternative to Solomonoff induction's universal distribution for general sequence prediction. I believe that Tom Sterkenburg told me that this is possible, but I could not find it written down anywhere (though I may have missed it!). Indeed, has been conjectured not to be limit=anytime computable by Hutter and Muchnik: https://arxiv.org/abs/cs/0407057 . I worked out the anytime algorithm with @Aram Ebtekar and @Marcus Hutter , though any mistakes are mine. Anytime computable (or limit computable): A function f is anytime computable if where is finitely computable. Lower semicomputable (or l.s.c.): A function f is l.s.c. if where is non-decreasing in t. Computable (or estimable): A function f is computable if where . A sequence predictor is a function from the binary strings to [0,1] which we interpret as the probability of seeing the prefix. Assuming "superadditivity" , specifies a (unique) distribution on possibly infinite sequences. The function is also called a semimeasure. Measures satisfy superadditivity with equality, which is called additivity. Solomonoff induction predicts with the universal distribution , which is lower semicomputable but (only) has anytime computable posteriors. is a u…

An anytime algorithm for mixing the computable measures

Epistemic status: Not peer reviewed, high chance of typos and small chance of errors. Written entirely by me, checked by Fable. In this post I prove the existence of an anytime computable Bayesian mixture of all computable measures called , and briefly argue that this is a reasonable alternative to Solomonoff induction's universal distribution for general sequence prediction. I believe that Tom Sterkenburg told me that this is possible, but I could not find it written down anywhere (though I may have missed it!). Indeed, has been conjectured not to be limit=anytime computable by Hutter and Muchnik: https://arxiv.org/abs/cs/0407057 . I worked out the anytime algorithm with @Aram Ebtekar and @Marcus Hutter , though any mistakes are mine. Anytime computable (or limit computable): A function f is anytime computable if where is finitely computable. Lower semicomputable (or l.s.c.): A function f is l.s.c. if where is non-decreasing in t. Computable (or estimable): A function f is computable if where . A sequence predictor is a function from the binary strings to [0,1] which we interpret as the probability of seeing the prefix. Assuming "superadditivity" , specifies a (unique) distribution on possibly infinite sequences. The function is also called a semimeasure. Measures satisfy superadditivity with equality, which is called additivity. Solomonoff induction predicts with the universal distribution , which is lower semicomputable but (only) has anytime computable posteriors. is a u…

Accel raises $550 Mn India fund, takes India corpus to $1.2 Bn in 18 months

US-based venture capital firm Accel has raised a $550 million early-stage fund for India as part of a $3.5 billion fundraise across four new funds globally. This is the first time Accel has raised all its global funds together. The new India fund is Accel’s ninth early-stage fund for the country. Accel has raised $1.2 billion for India over the past 18 months. This follows Accel’s $650 million early-stage fund for India and Southeast Asia announced in January 2025. The fund targets pre-seed to Series A startups across sectors such as AI, consumer, fintech and manufacturing. More than 50% of Accel’s eighth India fund is yet to be deployed, according to sources. The latest fund was also significantly oversubscribed, sources said. Accel does not expect to start deploying the new $550 million fund this year. According to sources, Accel has about $900 million in dry powder to deploy in India. Entrackr has reached out to Accel for a comment. The global fundraise includes a $1.35 billion growth fund that can back companies across geographies and stages. Accel recently used its previous growth vehicle to participate in the Series B round of fintech startup Sahi. In India, the firm will focus on AI across platforms, vertical applications, infrastructure and businesses that combine AI with human involvement. It will also continue to invest in consumer, manufacturing and fintech. Advanced manufacturing and deeptech are emerging areas of focus. Accel said it writes the first institution…

Wayland IME issue with Fcitx5 (Korean input) in ChatGPT Linux app

Hi, I found a reproducible IME issue with the new ChatGPT Linux preview. Environment: ChatGPT Linux: 26.803.81509 EndeavourOS (Arch Linux), x86_64 KDE Plasma / KWin Wayland Fcitx5 with Korean input Installed through the AUR package openai-chatgpt , which repackages the official OpenAI .deb binary Issue: When ChatGPT is launched normally, the Korean Hangul/English toggle key does not work inside the ChatGPT app. Fcitx5 itself is working normally, and the same Hangul/English toggle key works correctly in other applications on the same system. I found that launching ChatGPT with the following flags fixes the problem immediately: chatgpt --enable-features=UseOzonePlatform --ozone-platform=wayland --enable-wayland-ime After launching ChatGPT this way: the Hangul/English toggle key works normally Korean input works normally Steps to reproduce: Launch ChatGPT normally. Focus the message input field. Press the Hangul/English toggle key. The input method does not switch to Korean. Quit ChatGPT. Launch ChatGPT with: chatgpt --enable-features=UseOzonePlatform --ozone-platform=wayland --enable-wayland-ime Focus the message input field again. Press the same Hangul/English toggle key. Korean input now works correctly. Expected behavior: ChatGPT should work with the active Wayland/Fcitx5 input method when launched normally, without requiring additional command-line flags. I understand that EndeavourOS / Arch Linux is not currently an officially supported distribution. However, since the is…

Saber denies replacing Rideshare Stimulator’s writers with ChatGPT

After a former lead writer claimed Saber "replaced me with ChatGPT," CEO Matthew Karch now claims, "Neither Saber nor Unigine have replaced any writers with AI," for the Rideshare "Stimulator" game announced last month, developed by Unigine. The writer, Stella Sacco, says differently, however, posting on Bluesky that "I was lead writer on this one! […]

NVIDIA AI Factory Compute Is Becoming an Investable Asset Class

We announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent financing platforms designed to mobilize over $500 billion of third-party capital to support the buildout of AI infrastructure over time. This is a major milestone for NVIDIA and the AI industry. We have moved from an era in which companies […]

Drew Barrymore: Divorce took me years to get over

Drew Barrymore: Divorce took me years to get over

Pelosi: Those who knew of Biden’s decline ‘should have acted’

The former House speaker’s remarks came at POLITICO's The California Agenda: Sacramento Summit on Tuesday.