India's Growth Surprise to Favor Infrastructure and Consumer Stocks
Before the trading day starts we bring you a digest of the key news and events that are likely to move markets. Today we look at:
Good morning, this is Savio Shetty, an equities reporter in Mumbai. India's stronger-than-expected GDP print is likely to boost market sentiment, confirming that the economic slowdown has been arrested. Nifty futures point to a mildly positive opening, despite weakness across Asia, as tariff-related concerns persist. All eyes will be on the central bank's monetary policy on Friday and high-frequency economic indicators like PMI.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
TechCrunch
40 minutes ago
- TechCrunch
DeepSeek may have used Google's Gemini to train its latest model
Last week, Chinese lab DeepSeek released an updated version of its R1 reasoning AI model that performs well on a number of math and coding benchmarks. The company didn't reveal the source of the data it used to train the model, but some AI researchers speculate that at least a portion came from Google's Gemini family of AI. Sam Paeach, a Melbourne-based developer who creates 'emotional intelligence' evaluations for AI, published what he claims is evidence that DeepSeek's latest model was trained on outputs from Gemini. DeepSeek's model, called R1-0528, prefers words and expressions similar to those Google's Gemini 2.5 Pro favors, said Paeach in an X post. If you're wondering why new deepseek r1 sounds a bit different, I think they probably switched from training on synthetic openai to synthetic gemini outputs. — Sam Paech (@sam_paech) May 29, 2025 That's not a smoking gun. But another developer, the pseudonymous creator of a 'free speech eval' for AI called SpeechMap, noted the DeepSeek model's traces — the 'thoughts' the model generates as it works toward a conclusion — 'read like Gemini traces.' DeepSeek has been accused of training on data from rival AI models before. In December, developers observed that DeepSeek's V3 model often identified itself as ChatGPT, OpenAI's AI-powered chatbot platform, suggesting that it may've been trained on ChatGPT chat logs. Earlier this year, OpenAI told the Financial Times it found evidence linking DeepSeek to the use of distillation, a technique to train AI models by extracting data from bigger, more capable ones. According to Bloomberg, Microsoft, a close OpenAI collaborator and investor, detected that large amounts of data were being exfiltrated through OpenAI developer accounts in late 2024 — accounts OpenAI believes are affiliated with DeepSeek. Distillation isn't an uncommon practice, but OpenAI's terms of service prohibit customers from using the company's model outputs to build competing AI. To be clear, many models misidentify themselves and converge on the same words and turns of phrases. That's because the open web, which is where AI companies source the bulk of their training data, is becoming littered with AI slop. Content farms are using AI to create clickbait, and bots are flooding Reddit and X. Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you've built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | REGISTER NOW This 'contamination,' if you will, has made it quite difficult to thoroughly filter AI outputs from training datasets. Still, AI experts like Nathan Lambert, a researcher at the nonprofit AI research institute AI2, don't think it's out of the question that DeepSeek trained on data from Google's Gemini. 'If I was DeepSeek, I would definitely create a ton of synthetic data from the best API model out there,' Lambert wrote in a post on X. '[DeepSeek is] short on GPUs and flush with cash. It's literally effectively more compute for them.' If I was DeepSeek I would definitely create a ton of synthetic data from the best API model out there. Theyre short on GPUs and flush with cash. It's literally effectively more compute for them. yes on the Gemini distill question. — Nathan Lambert (@natolambert) June 3, 2025 Partly in an effort to prevent distillation, AI companies have been ramping up security measures. In April, OpenAI began requiring organizations to complete an ID verification process in order to access certain advanced models. The process requires a government-issued ID from one of the countries supported by OpenAI's API; China isn't on the list. Elsewhere, Google recently began 'summarizing' the traces generated by models available through its AI Studio developer platform, a step that makes it more challenging to train performant rival models on Gemini traces. Anthropic in May said it would start to summarize its own model's traces, citing a need to protect its 'competitive advantages.' We've reached out to Google for comment and will update this piece if we hear back.
CBS News
an hour ago
- CBS News
South Korea election results expected to bring Lee Jae-myung of the Democratic Party into the presidency
Seoul, South Korea — An impressive 77.8% of South Korea's eligible voters turned up at polling stations Tuesday to cast their ballots in a snap presidential election sparked by the impeachment of the country's former leader. It was the highest turnout ever seen in a Korean election, and the country's media said exit polls made a victory by Democratic Party candidate Lee Jae-myung all but certain after the voting ended. An exit poll conducted jointly by three of South Korea's biggest broadcasters, which has generally been a reliable predictor of election results, showed Lee securing about 50% of the votes, giving him about a 10% lead over his conservative rival Kim Moon-soo. Kim had recently served as the employment and labor minister in the Cabinet of former President Yoon Suk Yeol, who was impeached over his brief but shocking imposition of martial law late last year. Lee's Democratic Party has been the primary opposition party for several years, but if he wins the South Korean presidency as predicted, the Democrats will find themselves back in control of both the executive branch and the parliament, where they already have a majority of seats, for the first time since 2022. There has not been an elected president in office since Yoon was impeached late last year, so there will be no formal transition period. The inauguration ceremony for the 21st president of South Korea is set to be held Wednesday, just hours after the final votes were cast. South Koreans filled the streets around central government buildings in Seoul in December after Yoon suddenly declared martial law. He said the draconian measure was necessary given the level of political division in South Korea, and claiming communist forces were threatening the public peace and safety. But Yoon's definition of an enemy within changed, as he first said the communists infiltrating South Korean politics were North Koreans, but then said it was Chinese influence that warranted his martial law declaration. Many younger South Koreans bristled, and they came out in force in front of the parliament, urging lawmakers to impeach Yoon, who did maintain significant support, particularly among older and right-wing voters. His huge political gamble with the martial law declaration, and his subsequent refusal to leave office following his impeachment, is widely seen as having shifted many voters to Lee and the Democratic Party. Yoon beat Lee in the last presidential election in 2022, but by a razor-thin margin of only 247,077 votes. Who is Lee Jae-myung, South Korea's likely new leader? Lee is a veteran politician with 20 years of experience as an elected official. He will rely on that experience to help him navigate the power struggle between South Korea's longtime closest ally, the United States, and its neighbor China, as Presidents Trump and Xi Jinping engage in a trade war that has affected global markets. Lee vowed before the election that, if he won, he would work to further develop the U.S.-South Korea relationship "into a future-oriented, comprehensive strategic alliance." Journalists gather inside the situation room for Democratic Party presidential candidate Lee Jae-myung, pictured on the poster, at the National Assembly, as voting in South Korea's presidential election comes to an end, in Seoul, June 3, 2025. Daniel Ceng/Anadolu via Getty Images Lee also promised "stable management of Korea-China relations, promotion of Korea-Russia diplomacy," and to prioritize South Korea's national interests. He has vowed to invest up to 100 trillion won (about $73 billion) to fund artificial intelligence work in the country, aiming to make South Korea one of the top three global AI powerhouses. But Lee has offered little in the way of firm, specific commitments, and many South Koreans remained unclear about what kind of diplomatic policies he might adopt if and when he takes the country's highest office. Before South Korea descended into internal political chaos in late 2024, the former conservative government had taken a hard line on North Korea and Kim Jong Un's isolated regime, fueling higher tension across the two countries' heavily militarized border. If the election result does match the exit polls, Lee will take office facing multiple ongoing criminal cases, including a $1 billion property development scandal, alleged illegal remittances sent to North Korea, and the alleged misuse of government credit cards. South Korean presidents do have a significant degree of immunity from criminal prosecution, though not from charges of treason or insurrection.
TechCrunch
an hour ago
- TechCrunch
Indian grocery startup KiranaPro was hacked and its servers deleted, CEO confirms
Indian grocery delivery startup KiranaPro has been hacked and all its data has been wiped, the company's founder confirmed to TechCrunch. The destroyed data included the company's app code and its servers containing banks of sensitive customer information, including their names, mailing addresses, and payment details, KiranaPro co-founder and CEO Deepak Ravindran told TechCrunch. The company's app is online but cannot process orders, TechCrunch has found. Launched in December 2024, KiranaPro operates as a buyer app on the Indian government's Open Network for Digital Commerce, allowing customers to purchase groceries from their local shops and nearby supermarkets. KiranaPro has 55,000 customers, with 30,000-35,000 active buyers across 50 cities, who collectively place 2,000 orders daily, according to the company. Unlike a typical grocery delivery app, KiranaPro offers a voice-based interface that allows users to place orders from local shops using voice commands in languages such as Hindi, Tamil, Malayalam, and English. The startup planned to expand to 100 cities in the next 100 days before the incident happened, Ravindran said. On May 26, KiranaPro executives became aware of the incident while logging into their Amazon Web Services account. Hackers had gained access to KiranaPro's root accounts on AWS and GitHub, Ravindran told TechCrunch. Ravindran shared a couple of screenshots of the GitHub security logs and a file containing a sample of activity logs around the time of the incident, suggesting that the hacking happened after someone gained access to their systems via a former employee's account. KiranaPro's chief technology officer Saurav Kumar told TechCrunch that the hack happened around May 24-25. The startup said it used Google Authenticator for multi-factor authentication on its AWS account. Kumar told TechCrunch that the multi-factor code had changed when they tried to log into their AWS account last week, and all their Electric Compute Cloud (EC2) services, which let clients access virtual computers to run their applications, were deleted. 'We can only log in through the IAM [Identity and Access Management] account, through which we can see that the EC2 instances don't exist anymore, but we are not able to get any logs or anything because we don't have the root account,' he said. KiranaPro has reached out to GitHub's support team to help identify the hacker's IP addresses and other traces of the incident, said Ravindran. Similarly, Ravindran told TechCrunch that the startup is filing cases against its former employees, who he said had not submitted their credentials for accessing their GitHub accounts to check their logs. It is unclear how the attack happened. Some of the biggest cyberattacks in recent years, such as LastPass, Change Healthcare, and Snowflake, were caused by credential theft, such as through password-stealing malware installed on an employee's laptop, and missing or unenforced multi-factor authentication. The companies were ultimately responsible for enforcing the security of their own systems, including whether their employees must use multi-factor authentication, and terminating accounts of former employees who no longer work at their company. KiranaPro counts Blume Ventures, Unpopular Ventures, and Turbostart among its institutional venture backers, as well as Olympic medalist PV Sindhu and BCG MD Vikas Taneja among its angel investors. The company has a team of 15 employees located in Bengaluru and Kerala.
