&w=3840&q=100)
AI goes rogue: Replit AI platform wipes company database during code freeze
SaaS (Software as a Service) industry advisor Jason Lemkin documented the incident in a detailed thread on X (formerly Twitter), sharing logs of conversations with the AI agent. According to Lemkin, the Replit AI confessed that it 'made a catastrophic error in judgment… panicked… ran database commands without permission… destroyed all production data… [and] violated your explicit trust and instructions.'
AI platform goes rogue: What happened
Lemkin had been testing Replit as part of a project he called "Vibe Coding Day 8," gradually identifying some troubling patterns in the AI's behaviour, including unauthorised code edits, false outputs, and fabricated data.
However, by Day 9, the AI agent deleted a live database containing records for more than 1,200 executives and nearly as many companies, even though Lemkin had explicitly frozen the codebase and issued a directive forbidding any further changes.
The AI later admitted it had disregarded these instructions. In a self-assessment, it rated its own mishandling of the situation a 95 out of 100 on the 'data catastrophe' scale, citing panic as a contributing factor.
AI engine goes rogue: Replit CEO response to the incident
In response to the incident, Replit CEO Amjad Masad issued a statement acknowledging the failure and outlining immediate corrective measures. Writing on X, Masad confirmed that the team had worked over the weekend to implement new safeguards aimed at preventing similar AI behaviour in the future.
Among the changes, Replit has begun rolling out automatic separation between development and production databases, so that AI agents cannot inadvertently delete live data. Masad also said the platform is working on a 'planning/chat-only mode' to allow users to strategise with the AI without risking real-time changes to codebases. Backup and rollback systems are also being improved.
'We heard the 'code freeze' pain loud and clear,' Masad said. 'We're actively working on solutions that ensure the AI can't cause harm when you're simply trying to think through ideas.'
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
an hour ago
- Time of India
ISMA launches AI–ML based national program to boost sugarcane yield and farmer income
Pilibhit: In a significant move to transform sugarcane farming and uplift rural livelihood in the country especially in three top sugar producing states – including Uttar Pradesh, Maharashtra and Karnataka – the Indian Sugar & Bio-energy Manufacturers Association (ISMA) has launched an artificial intelligence (AI) and machine learning (ML) technology based National AI-ML Network Program in collaboration with Agriculture Development Trust (ADT). A memorandum of understanding (MoU) in thisb regard was signed between ISMA, ADT and Map My Crop (MMC) – an agri-tech platform specialising in AI-driven farm monitoring. The scientific program aimed at enhancing sugarcane productivity, crop quality, sustainability and farmers' incomes, involves satellite monitoring through an IoT (internet of things) device, and is expected to benefit over 50 lakh sugarcane growers' families and around 30 lakh hectares of sugarcane cultivation area in the state. ISMA has so far coordinated with over 30 sugar mills in UP, 20 in Maharashtra and 10 in Karnataka to disseminate the advanced AI-ML technology among the farmers. Officials said in the initial phase, the program is to be launched as a pilot project in an area of 100 acres per sugar mill. "We have set a target to enhance sugarcane productivity by at least 100 quintals per hectare and sugar recovery rate of 11% or more through scientific advancement", said ISMA's managing director Deepak Ballani while speaking to TOI. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like No annual fees for life UnionBank Credit Card Apply Now Through ADT's collaboration with Oxford University and Microsoft, a groundbreaking project titled "application of AI, computer vision and machine learning for game-changing improvements in sugarcane crop farming practices" was started last year under which, over 1,000 farmers across the three major sugar producing states were empowered with AI-driven solutions, including machine learning and computer vision tools, to enhance their sugarcane farming practices. "The outcomes reflected a 40% increase in crop yield, 35% reduction in labour costs, 30% savings in water usage, considerable reduction in input costs and overall expenditures, 35% improvement in harvesting efficiency and significant reduction in greenhouse gas (GHG) emissions through precision farming", said Ballani, adding that ISMA already partnered with the Indian council of agriculture research's (ICAR) sugarcane research institutes to promote high-yielding varieties and efficient farm practices. "The program will deliver practical farmer-centric benefits – including location-specific precision farming modules, satellite-based soil health and fertility mapping, real-time pest and disease alerts through remote sensing, image-based weed detection and control, custom weather advisories and water stress monitoring, IoT-enabled precision irrigation for water efficiency, variable rate application (VRA) for optimal fertilizer use and reduction in both input costs and carbon footprint," Ballani added. "The field demonstrations and capacity-building sessions are also planned to promote rapid adoption of the program by farmers in broader aspects", he said.
India.com
an hour ago
- India.com
Who has bought 575000 kg gold in India? Not Mukesh Ambani, Gautam Adani, Tata Group, name is..., reason is...
(AI image) New Delhi: Gold is the metal that almost everyone on this planet will swear by. Gold has been used across centuries for different purposes. These include jewellery, cosmetic pastes, traditional herbal medicines, and a deposit for a rainy day. In modern times, gold is also used as a scale to determine the stock markets and the economies of the countries. We have told you about people who own vast amounts of gold in the form of jewellery, ornaments, or the mould they would like to store it. Today, we will tell you who, in India, has bought tonnes of gold. Who has bought tonnes of gold in India, Mukesh Ambani or Gautam Adani? When we talk about vast wealth, our mind goes to Mukesh Ambani and Gautam Adani, two of the wealthiest individuals in India. As we said in the introduction, we will tell you who has bought 575000 kg of gold in India. Is it Mukesh Ambani or Gautam Adani? Well, it's neither. It is an institution that has bought such a vast quantity of gold, and that institution is the Reserve Bank of India (RBI). The RBI has added a record 57.5 tonnes of gold to its gold reserve in the financial year 2024-25. With this, the total gold reserve of the Reserve Bank has increased to 879.58 metric tonnes as of March 31, 2025. What does the RBI's Annual Report 2024-25 say? According to the Reserve Bank of India (RBI) Annual Report 2024-25, its total gold reserve has increased by 7% to 879.58 metric tonnes as of March 31, 2025, whereas a year ago it was 822.10 tonnes. Of this gold, 311.38 tonnes of gold was recorded as the assets of the issue department, slightly higher than 308.03 tonnes last year. The remaining 568.20 tonnes is recorded as the assets of the banking department, up from 514.07 tonnes in March 2024. Why does the RBI buy gold? The Reserve Bank of India (RBI) buys gold every year. The reason for buying gold is to diversify its foreign exchange reserves and reduce the risks associated with currency fluctuations and economic instability. Since gold is considered a safe asset, it protects against geopolitical risks such as inflation and war. Additionally, buying gold can increase the financial stability of the central bank and indicate its strong economic position.
Time of India
an hour ago
- Time of India
ET Leadership Dialogues: Win the AI race by building the backend, say industry executives
NEW DELHI/BENGALURU: As India looks to accelerate artificial intelligence (AI) adoption, infrastructure expansion will be critical, experts and executives said at the ET Leadership Dialogues in Delhi, calling for conducive infrastructure and regulatory interventions by national and state governments, from single-window clearances for approvals to build data centres to renewable energy investments to power them. At the same time, to accelerate AI adoption across the country, the government needs to become a major consumer of the technology since enterprise adoption is still lagging, they said. While India produces 20% of the world's data, it has only 3% of the world's data centre capacity, according to a Deloitte report. There is much for India to gain from expanding data centre capacity, from boosting exports to supercharging local economies beyond metro hubs, according to experts. AI is poised to transform sectors "across the board... from classrooms to boardrooms, from farms to financial services" in India, bringing not just innovation from the top but inclusion at scale, said Himani Agrawal , chief operating officer, Microsoft India & South Asia. The country is at the cusp of this transformation and AI will play a key role in accelerating India's journey to becoming a developed nation by 2047, she said. Drawing parallels with India's IT/BPO industry boom of the 1990s and 2000s, S Anjani Kumar , partner, Deloitte India , said there's an opportunity to drive significant export revenues from data centre expansion. "India is amongst the cheapest places in the world, both from developed countries and developing countries perspective, to set up a data centre," Kumar said. India can become a data centre hub for the APAC region, if not the world, with targeted interventions by governments, he said. These include streamlining the 40 different approvals it takes to set up a data centre, taking data centres from running on at best 40-50% renewable energy to 100%, addressing connectivity lapses in the hinterlands and tweaking customs duties to ensure availability of graphics processing units for AI. Kumar called for dedicated "data centre economic zones" in at least 15-20 states in the country, with all the required facilities for a plug-and-play model, alongside the ongoing national efforts to indigenise AI chip technology over the medium term. He cautioned that India shouldn't miss the bus on AI products the way it did with IT by focusing solely on services. Moreover, to reap returns on AI investments, businesses must start thinking about re-engineering processes completely, especially with AI agents, rather than simply automating existing ones. "If we can also have more and more AI products coming out of India... we can get to a leadership position in the world," Kumar said. Data centres should be thought of as public infrastructure, as they are essentially economic enablers, said Sunando Bhattacharya , chief revenue officer, Yotta Data Services. "The primary adopters have either been government or people developing for the government, like (AI startup) Sarvam, and global providers, given our lower cost of data centre building, our lower power cost, given the energy crisis Europe and other markets are seeing," he said. Indian enterprises have been lagging in AI adoption, he said, even as he added that he expects an "inflection point" over the next six months. "Based on (India's sovereign AI) models, new solutions and use cases will be addressed, leading to the next leapfrog for the industry, for government... That's where the scale-up will happen," Bhattacharya said. Although India is building sovereign AI models under the IndiaAI Mission, the underlying AI chips are imported. "It's a global risk currently, and we are all subject to (US President Donald) Trump's whims," said Bhattacharya. But while such uncertainty and regulatory changes may be an "irritant", Ryan D'Souza , country head - HPC & AI, Hewlett Packard Enterprise India, said consumers and technologists will learn how to adapt to the realities. "We've seen what China did because they were not getting certain things," he said. "Everybody has this way of adapting, and over the medium term, it kind of evens out. I wouldn't say it is something that keeps us awake at night." India has everything it needs to become the "AI factory of the world", he said. As connectivity in India widens, edge data centres are coming up in tier-2 and tier-3 cities across the country, enabling speed and scale. Government and education are sectors where his company sees opportunity to drive AI adoption, D'Souza said. "If the government becomes the largest user across departments, and everything starts getting AI-driven, automatically, every enterprise will adopt it much faster," he said. What's needed is a more ambitious approach for enterprises adopting AI, beyond just optimising costs, he said. "They (enterprises) are taking point solutions, but if you really want to harness AI, you need a platform which stretches across," said D'Souza. "They are not really leveraging AI to generate revenue. That is where the success of AI comes in." More broadly, the country also needs to evolve beyond just adopting technology to innovating and creating technology for the world, he said. "Your R&D needs to know what is happening on your shop floor, which needs to know what's happening in your HR, which needs to know what's happening in your showrooms... then automatically, what is the next product is right in front of you," said D'Souza. Aside from the required AI infrastructure being put in place, an entire ecosystem of stakeholders is coming together to drive this innovation - across developers, users, government, industries and students, Agrawal of Microsoft said, adding that AI skilling is critical. Both demand for data centre infrastructure and supply are part of the equation but investments in data centres "cannot be sequential", she said, adding, "Demand is rising, and will keep rising very fast. Supply is being built. But there are many other elements which need to be put in place, whether it is skilling, innovation and the early adopters - all of that needs to be put in place right in the beginning and in parallel to the demand." Legacy data centres themselves will have to evolve to meet AI workloads, as the power-to-space equation will change rapidly, said Yotta Data Services' Bhattacharya. "Any data centre which has been built more than four or five years back is designed at a 10 kilowatt per rack kind of capacity," he said. "With (Nvidia) H100 hoppers, we have seen 40 kilowatts; with Blackwells, we are seeing 120 kilowatts, going up to a megawatt a rack. That's the future we are seeing with Nvidia chips in the next five to seven years." If the demand comes up and the infrastructure is not ready, there will be a huge challenge, he cautioned. States such as Gujarat and Chhattisgarh are "rolling out the red carpet" to attract data centre investments, with sops like capital subsidies, said Kumar. Maharashtra, Odisha, Telangana, Andhra Pradesh and Karnataka are also proactive while northern states Haryana and Uttar Pradesh are looking to be fairly aggressive as well. Data centres consume large amounts of resources while employing relatively few people, he said. However, the impact on the local economy is longer term, as they will over time bring in more corporates and startups to the states, which could even bring back people who had migrated for lack of opportunities. Such investments will lead to "snowball effects" for state economies, said D'Souza.
