Cyberattacks are an unfortunate problem for digital business, targeting small companies to the largest enterprises. As digital infrastructure expands and more sensitive information is stored online, security risk management needs must go beyond prevention to ensure the organization has full visibility of their digital environments and can address incidents in real time. Join our upcoming summit to learn A Holistic Approach to Endpoint Detection and Response to get practical EDR strategies to bolster your security strategy. You’ll learn how the principle of least privilege, IoT security and telemetry helps protect your endpoints, and receive advice for using advanced forensics and AI-powered investigations to speed response times.
As cloud use continues to expand, proper management, monitoring and security is critical to ensure organizations are gaining the most benefit. Organizations navigating digital transformation initiatives know the cloud is a major playing ground for emerging technologies related to cloud migration, cloud security and containerization. Tune into the Managing Hybrid and Multi Cloud Environments Summit to hear leading experts in the field discuss the latest strategies for monitoring complex cloud environments, preventing cyber attacks and security breaches via the cloud, and how to integrate the latest technology trends into cloud usage.
Top Stories
Brain science: Mind mechanism of AI safety, interpretability and regulation May 28, 2024 by David Stephen The basis of how the human brain works is conceptually the mechanism of the mind—which is the electrical and chemical signals of neurons, in sets, with their interactions and features. Recently, the Department of Commerce released a Strategic Vision on AI Safety, stating that, “The U.S. AI Safety Institute will focus on three key goals: Advance the science of AI safety; Articulate, demonstrate, and disseminate the practices of AI safety; and Support institutions, communities, and coordination around AI safety.”
A Look at Kwaai’s Personal AI OS May 28, 2024 by Dan Wilson In the latest episode of the AI Think Tank podcast, I had the privilege of hosting Toby Morning and Karsten Wade from Kwaai. These two trailblazers are at the forefront of a nonprofit initiative dedicated to democratizing AI through the development of a Personal AI Operating System (pAIOS). Our discussion spanned various topics, from the origins of Kwaai to the technical intricacies of pAIOS, providing a comprehensive insight into their groundbreaking work. This article delves into the key points and takeaways from our enlightening conversation.
Becoming an AI Utility Function Exercise – Part 2 May 26, 2024 by Bill Schmarzo The media, including movies, have made AI conversations confusing and overly emotional. I propose a simple exercise for middle and high school students to provide hands-on training in understanding how an AI model works and the vital role of the AI Utility Function. The exercise will focus on something we do nearly every day: deciding where to dine. We will follow a four-step process to define, create, and execute our “Dining Recommendation” AI Utility Function.
In-Depth
Enhancing data lineage and metadata management in ELT pipelines May 28, 2024 by Ovais Naseem ELT pipelines facilitate the seamless movement of data from source systems to target destinations, enabling transformation and analysis along the way. However, as data traverses through these pipelines, maintaining visibility into its lineage and managing metadata becomes paramount for ensuring data quality, compliance, and governance.
Contextualize your business data with content orchestration techniques May 23, 2024 by Alan Morrison One of the major issues enterprises have is tapping into business information that’s trapped in many different siloed applications. Customer data platforms (CDPs) are supposed to unify structured data about customers from a number of these siloed applications and make that information accessible to a broad range of users. But what about content? Textual content because it behaves differently in digital form is disconnected from images and video, which are in turn disconnected from customer transactional data.
Cognitive robotics – Part one May 22, 2024 by Ajit Jaokar In this three-part series, we will explore cognitive robotics – a fascinating subject that promises to play a major role in the evolution of AI. Cognitive robotics lies at the intersection of robotics, artificial intelligence (AI), and cognitive science, aiming to create intelligent systems that mimic human cognitive processes.
Data management demand in digital twin-oriented intensive care units May 22, 2024 by Alan Morrison One of the more telling exhibits at April’s NTT Upgrade 2024 event in San Francisco was the Autonomous Closed-loop Intervention System (ACIS) in development at NTT Research’s Medical & Health Informatics Lab (MEI). The ACIS concept underscores how hospitals can embrace a systems-level automation initiative starting with the intensive care unit to create a patient-centric, personalized, continually updated data management environment that can radically boost patient outcomes.
DSC Weekly 21 May 2024 May 21, 2024 by Scott Thompson Read more of the top articles from the Data Science Central community.
As business data demands increase, cloud providers and their customers find themselves having to consider the implications of increasing storage costs, security risks and environmental footprint. Such impacts are of particular importance to U.K. organisations, as it is the largest cloud market in Europe.
TechRepublic spoke to U.K. cloud experts to identify the top five industry trends emerging from the country’s burgeoning reliance on the fundamental IT infrastructure. These cloud trends are:
Premiumisation of cloud packages.
Movement towards hybrid multicloud models.
Influx of sustainable cloud solutions.
Continuous quest for data sovereignty.
Focus on cloud security.
1. Premiumisation of cloud packages
Adrian Bradley, the head of cloud transformation at KPMG U.K., explained the rising costs of cloud providers’ most premium services are forcing companies to choose their packages more carefully. According to research from Dark Matter, 90% of U.K. businesses have noted their cloud costs rising.
The reduction in demand for best-in-class offerings is also pushing cloud providers to improve their ability to switch from public cloud and diversify the bundles they offer. Bradley referred to this concept as premiumisation.
He told TechRepublic in an email, “Rising public cloud prices are mostly a result of high energy prices, a shortage of computer chips and increased demand caused by the growing use of generative AI. While expensive, the premium services provided by public cloud are extremely valuable, as organisations use them to drive maximum efficiency into their enterprises.
“But where enterprises cannot get that value (from the premium services), some workloads can be placed elsewhere more economically. Non-public cloud offerings are proving to be an attractive alternative, providing a cost-effective option for less premium services.
“As a result, UK businesses will need to review their enterprise and cloud strategies to become more adaptable and value-oriented. This means that their use of public cloud will focus on higher-value, premium services like generative AI, which will enable more complex and intelligent solutions for data analysis, automation and decision-making. Simple storage and compute will gravitate to the lowest-cost platform.”
2. Movement towards hybrid multicloud models
According to a December 2023 Enterprise Cloud Index from cloud platform provider Nutanix and reported on Cloud Next, 46% of U.K. businesses are set to utilise multiple public clouds in the next one to three years, while globally this figure is predicted to be just 26%. Hybrid multicloud models are also set to be used by 26% of U.K. businesses, compared to 19% today. The main factors cited by U.K. respondents behind this pronounced shift towards hybrid multicloud models are performance, cost, data sovereignty, malware protection and flexibility.
Jake Madders, the co-founder of U.K.-based cloud hosting provider Hyve, says that partnering with different cloud providers is more cost-effective as the price of services increase. He told TechRepublic, “Companies can optimise their expenditure based on workload requirements and price differences among providers, thereby reducing total cloud expenses.”
The issues associated with vendor lock-in are also becoming more apparent. In April 2024, documents seen by The Register revealed the U.K. government was concerned its current cloud model, dominated by AWS and Azure, put its “negotiating power over the cloud vendors” at risk. Distributing workloads across multiple providers can reduce such risks, as well as the potential impacts of outages and data breaches.
Madders added, “This type of cloud infrastructure also allows for greater resilience and performance by providing redundancy and enabling workload distribution across geographically dispersed data centres. This ensures high availability and minimises latency for improved user experience.”
3. Influx of sustainable cloud solutions
“Businesses today do not just want power; they want sustainable power and accessibility,” Lars Nyman, chief marketing officer of U.K.-based cloud computing platform CUDO Compute, told TechRepublic in an email. “They also want to contribute to a greener future while not losing out on high-performance computing.”
While the term “cloud” may conjure images of fluffy white puffs in blue skies, the reality is the technology is not inherently environmentally friendly. Many data centres are still reliant on fossil fuels, while the applications and databases hosted there are not optimised to use the resources efficiently. Research from Intel predicts infrastructure and software inefficiency count for more than 50% of greenhouse gas emissions in the data centre.
Madders added, “Environmental concerns around hosting and data centres are still one of the major technology drivers in the cloud industry. As a result, we are likely to see new cutting-edge technology in cooling systems and computing power.”
Such new technologies might include energy-efficient liquid cooling systems and processors that use dynamic voltage and frequency scaling. There could also be developments towards reusing the excess heat from data centres; the U.K. government recently announced it would channel it to provide low-cost heating for more than 10,000 homes.
Nyman added that these new technologies could work to democratise sustainability in the area. “Previously, only large enterprises could afford to pursue meaningful sustainability goals,” he told TechRepublic. “Startups (needed) to focus on keeping the lights on.
“Dirty energy-guzzling data centres may eventually become a thing of the past.”
4. Continuous quest for data sovereignty
Jason Van der Schyff, chief operating officer at London-based private cloud provider SoftIron, told TechRepublic in an email, “We see little to suggest that 2024 will be any less turbulent in terms of geopolitics than we have seen in years past.” Earlier this month, the payroll system used by the Ministry of Defence was hacked, and ministers reportedly suspect the involvement of China.
“With regard to its impact on IT, we expect that we will see this accelerate plans by nation-states to boost their own sovereign resilience,” Van der Schyff added. He predicted this will manifest as governments investing in infrastructure and IT skills to build out “true sovereign clouds.” In January 2024, The Times reported that the U.K. government would support the growth of the country’s data centre infrastructure. Then in March, the government announced it would invest more than £1.1 billion to train in AI, quantum and other future tech.
Prakash Pattni, the managing director of Financial Services Digital Transformation at IBM Cloud, says organisations will take meaningful steps towards achieving their own data sovereignty to aid them in compliance with new regulations.
He told TechRepublic in an email, “As regulations evolve, enterprises are finding that they need to be prepared to navigate geographic-specific requirements to remain competitive and the cloud can play a pivotal role in helping enterprises to achieve data sovereignty.
“This is especially critical now as AI grows – and with it – comes an influx of data. While AI will fuel tremendous business innovations, it also requires strategic considerations around where data resides, data privacy and more.
“Organisations throughout the U.K., and especially those in highly regulated industries, are embracing sovereign cloud capabilities to help them manage their regulatory obligations and will continue to do so in the coming years.”
5. Focus on cloud security
Neil Templeton, the senior vice president of network-as-a-service platform provider Console Connect, told TechRepublic in an email, “Cyberattacks are inevitable, and their frequency will only increase, especially as hackers employ AI to boost their efforts.” In January 2024, the U.K.’s National Cyber Security Centre ruled that generative AI may increase the risk of cyber threats as it provides “capability uplift.”
Templeton added, “Network security and infrastructure should be a top priority this year, and part of the assessment should be to determine if businesses should avoid the risks of the public internet by moving to a private network environment.”
IBM Cloud’s Pattni added that, this year, many U.K. companies are prioritising their cyber security when it comes to their cloud services. He said, “Enterprises across highly regulated industries dealing with sensitive data – such as healthcare, telco, financial services and the public sector – are increasingly adopting risk management solutions that can help them gain visibility across their entire IT estate including third and fourth parties.”
“It’s critical that enterprises have the right foundation in place to truly enable trusted performance and security for enterprise AI and other data-intensive workloads.”
Artificial Intelligence has become one of the fastest-growing fields of our recent years. These new technologies will change the world forever, presenting us with a significant choice:
Embrace AI to optimise our workflow or ignore it altogether.
If you're reading this, I’m pretty sure you fall into the first camp: you want to take advantage of AI to enhance your daily tasks.
However, there’s a significant problem. Major companies are striving to distinguish themselves by flooding the market with a diverse range of applications. This makes it hard to know all the options that are out there waiting to be used.
This article aims to introduce Google's top AI applications and guide you on how to start using them today.
Google Gemini
Gemini is an advanced Large Language Model (LLM), a direct competitor of OpenAI’s GPT models, available for everyone to use for free!
Screenshot of Gemini main interface
It integrates Google's extensive research in AI to provide more accurate, context-aware responses, making it highly effective for applications like chatbots, virtual assistants, and other language-based AI solutions. Some of the key characteristics that make Gemini one of the best LLMs in the market are:
Free to Use: No payment is required.
Multimodal Capabilities: It can deal with images and documents.
User-Friendly Interface: Easy to use.
Getting Started
To start using Google’s Gemini, you just need to have a Google account. You can go to Gemini’s main interface and start chatting with it today.
Google Cloud
Set of cloud computing services that run on Google’s infrastructure. It offers multiple services including computing, storage, analytics, and machine learning. It can be used to build, train, and deploy AI models and solutions.
Its main AI services include:
AutoML: Designed with ready-to-use APIs for users with basic ML knowledge. Its main tools include AutoML Vision or AutoML Natural Language among others.
Vertex AI: A managed platform that simplifies the ML lifecycle, from data preparation to model deployment and monitoring.
Natural Language API: An intuitive tool to deal with natural language, supporting sentiment analysis, entity recognition, and syntax analysis.
Getting Started
To start using it today you just need a Google Cloud account. If you are a new user, you will get 300$ credit as a free trial during the first 90 days. Once your account is created, you can access these tools via the Google Cloud Console.
TensorFlow
TensorFlow is an open-source framework to create from scratch and train ML models. It is compatible with multiple programming languages such as Python and Java.
TensorFlow's main features include:
Comprehensive Ecosystem: TensorFlow offers tools to support the whole ML project lifecycle: model building, training, and deployment.including
Community: It is a widely used tool with extensive documentation and an active community.
Cross-Platform: With TensorFlow Lite, models can be deployed on a variety of devices, from smartphones to IoT devices.
Getting Started
Install TensorFlow using pip:
pip install tensorflow
Then you can explore the wide range of tutorials available on the TensorFlow website.
Experiments with Google
It is a collection of AI projects and experiments. It provides an interactive way to explore what AI can do, with many projects available for direct interaction. It uses hardware components and software tools from Google.
Main features include:
Educational Tools: Most of the projects are crafted to teach basic AI and ML concepts interactively and funnily.
Interactive Demos: You can interact with AI models and see how they respond to different inputs.
Getting Started
Visit Experiments with Google to explore the available projects. Many of these experiments include code and instructions for those interested in the technical details.
AI Hub
Google AI Hub is a platform where you can share and deploy ML models and pipelines. Its main goal is to foster a collaborative synergy between users and enterprises to leverage pre-build models and solutions.
Its main features are:
Pre-trained Models: Access a library of models pre-trained by Google and other users.
Collaboration: Share your models and solutions with the community, and discover other users’ models.
Integration with Google Cloud: Seamlessly integrate with other Google Cloud services for deployment and monitoring.
Getting Started
AI Hub is part of Google Cloud, so you'll need a Google Cloud account to access it.
Browse the AI Hub for models and solutions that fit your needs.
Wrapping Up
Google offers a wide variety of tools to start your journey into AI and machine learning. By exploring and taking advantage of them, you can optimize your workflow, enhance your daily tasks, and stay ahead in this fast-paced field!
Josep Ferrer is an analytics engineer from Barcelona. He graduated in physics engineering and is currently working in the data science field applied to human mobility. He is a part-time content creator focused on data science and technology. Josep writes on all things AI, covering the application of the ongoing explosion in the field.
More On This Topic
Exploring the Zephyr 7B: A Comprehensive Guide to the Latest Large…
Exploring the Latest Trends in AI/DL: From Metaverse to Quantum Computing
KDnuggets™ News 22:n01, Jan 5: 3 Tools to Track and Visualize…
A Beginner's Guide to End to End Machine Learning
Essential Machine Learning Algorithms: A Beginner's Guide
In light of criticism over its approach to AI safety, OpenAI has formed a new committee to oversee “critical” safety and security decisions related to the company’s projects and operations. But, in a move that’s sure to raise the ire of ethicists, OpenAI’s chosen to staff the committee with company insiders — including Sam Altman, […]
If you've used Google Search over the past several weeks, you've probably noticed a new feature. Instead of getting right to relevant links (or paid advertisements) when you search for something, the search engine has started displaying an AI-generated summary that, in some cases, might have the answer you're looking for.
To generate these summaries, Google combines answers from separate websites into a summary "overview" that addresses your query. Google says the AI-powered summaries only appear when the "responses can be especially helpful." But in my experience, I've found them to be flat-out wrong often enough that I scroll right past them.
Also: 5 useful AI features Google just unveiled for Chromebook Plus
I'm not alone.
The official Google support forum has been filled with posts from users asking how to turn off the Google AI responses — either because they get in the way of the results Googlers are looking for or because the information offered up is bad. Google quickly locks those posts, though, and comments are disabled.
So, what can you do if you don't want to deal with Google AI overviews?
Like the much-maligned Meta AI, there's no official way to turn it off. But, also like Meta AI, there are some workarounds you can try.
How to avoid Google Search AI summaries: Three ways
Web view
The first way around Google's AI summaries involves a new tab at the top of search results. When you search for something, you'll see tabs above the results that let you limit that search to images, videos, news, shopping, and so on. But a new tab has been added labeled "Web." Clicking that takes you straight to web search results for your query.
Also: How to sign up for Google Labs — and 5 reasons why you should
This does involve an extra step, but it's an effective way to avoid having AI try to answer your question.
2. Extensions
Second, several developers have created Chrome extensions to shut the door on Google AI butting in. You can't install extensions on the Android version of Chrome, so this is a desktop-only solution.
Because Google search is an HTML page, and any part of an HTML page can be removed by the browser, these extensions simply turn off that part of the page. You can find these by going to the settings menu at the top right, hovering over "Extensions," and choosing "Visit Chrome Web Store."
The extensions are listed under names like "Hide AI Overviews" and "Bye Bye, Google AI."
3. Use this search shortcut
Lastly, you can set your browser to automatically default to the aforementioned "web view" every time. If you're using Chrome, just right-click in the address bar and choose "manage search engines." Scroll down until you see "Add." At this point, you're creating a new search shortcut. The name doesn't matter, just call it something you'll remember. The same goes for the shortcut. But for the URL, use https://www.google.com/search?q=%s&udm=14 . Once you do that, your Google searches will automatically default to web view, at least when you search from the address bar.
Google likely won't provide an official way to opt out of this feature for users. Companies can't opt out of appearing in AI overview either. On a page that answers the question "How to prevent content from appearing in AI Overviews," Google's non-answer is simply "AI Overviews offer a preview of a topic or query based on a variety of sources, including web sources. As such, they are subject to Search's preview controls."
CoreWeave, a cloud computing startup that specialises in providing GPU resources, has seemingly broken into the ranks of hyperscalers by raising a staggering $7.5 billion in debt financing from investors, including Blackstone and Magnetar, in one of the largest-ever private debt financings.
In addition to this, the company recently announced a $1.1 billion Series C funding round led by Coatue, with participation from Magnetar, Altimeter Capital, Fidelity Management & Research Company, and Lykos Global Management.
With this, CoreWeave has raised $12 billion in equity and debt investments over just as many months. It plans to use the funds to scale its data centre capacity to 300 megawatts, doubling the number of centres from 14 to 28 by year-end.
Coreweave co-founder and CSO Brannin McBee recently mentioned that the company has already secured power and data centre space to more than double that amount in 2025.
This, coupled with the fact that they grew from three to 14 data centres and quadrupled their employees over the past year, speaks highly of the pace of their growth.
McBee also mentioned that another debt deal could happen this year, further fueling the company’s growth.
McBee believes the funding rounds have propelled CoreWeave into the ranks of hyperscalers, alongside industry giants like Amazon, Google, Microsoft, and Oracle.
However, he emphasised that CoreWeave stands out as an “AI hyperscaler”, differentiating itself through its specialised focus on AI infrastructure.
US-based Coreweave, which started out as a GPU provider to crypto-miners in 2017, witnessed its revenue skyrocket after the pivot.
While the company lowered its projected revenue for the year from $630 million to just over $500 million, it still registered a revenue of $440 million in 2023, an increase of more than 17-fold or 1,660% from $25 million in 2022.
Looking ahead, they are even more optimistic, with projections of $2.3 billion in revenue for 2024 based on having over $7 billion in signed cloud contracts through 2026.
This rapid growth has attracted the attention of major investors, including NVIDIA, which has taken an equity stake in the company.
Competing with Cloud Giants
Despite the impressive revenue, the company still has a long way to go to catch up with the established cloud giants.
In Q1 of 2024, AWS and Google Cloud generated $25 billion and $9.6 billion respectively. On the other hand, Microsoft’s Intelligent Cloud group, which includes Azure, reported a revenue of $26.7 billion.
However, CoreWeave’s focus on AI workloads, alongside its close partnership with NVIDIA and access to its latest AI chips, have given it a competitive edge in this fast-growing market segment.
NVIDIA is using its market dominance in GPUs to build up CoreWeave as a counterbalance to the cloud computing giants. The giant which supplies a majority of CoreWeave’s chips, views this trend favourably, possibly for reasons related to leverage.
It is reported that NVIDIA has granted preferential access to its GPUs to certain alternative cloud providers.
NVIDIA allocated a significant portion of its limited H100 supply to CoreWeave and other smaller cloud providers, while larger players like AWS and Microsoft Azure struggled to secure adequate supplies.
This allowed CoreWeave to attract major customers, including Microsoft itself, which signed a deal to rent GPU servers from CoreWeave to support its own Azure cloud customers.
This would also ensure sufficient computational power for OpenAI to train its generative AI models.
CoreWeave boasts an impressive GPU infrastructure, delivering access to over 45,000 GPUs and offering the industry’s broadest range of NVIDIA GPUs with 13 SKUs available on demand.
The company has deployed the largest fleet of NVIDIA A40 GPUs in North America and provides a variety of NVIDIA GPU types purpose-built for different use cases.
This includes the Quadro RTX series and NVIDIA Ampere architecture GPUs like the A100 for PCIe (80GB), a40, H100 HGX (80GB), L40, L40S, GH200, B200, empowering businesses to harness the latest advancements in GPU technology for their computing needs.
CoreWeave also expects to have the availability of NVIDIA’s new GB200 chip in early 2025, addressing the immense demand for this cutting-edge technology.
The exact number of GPUs CoreWeave has in comparison to hyperscalers like Amazon, Google, and Microsoft might be lower. However, CoreWeave emphasises a key difference in its approach. It offers a much broader range of NVIDIA GPU types, contrasted with the “one size fits all” approach of large, generalised clouds. Furthermore, CoreWeave claims to deliver a performance-adjusted cost of up to 80% less expensive than its competitors.
For instance, Coreweave charges $2.39 per hour to rent an NVIDIA A100 40GB GPU, which is commonly used for model training and inference. This translates to a monthly cost of $1,200.
In contrast, the same GPU on Azure costs $3.40 per hour, or $2,482 per month, while on Google Cloud, it costs $3.67 per hour or $2,682 per month.
CoreWeave’s Edge
Despite being much smaller, the company, in its investor pitch has drawn direct comparisons to major cloud providers, particularly AWS.
The company’s documents have previously projected a gross margin of 85%, derived from the difference between operational costs and rental revenue.
CoreWeave’s 2023 revenue is also higher than that of some competitors in the specialised GPU cloud market, such as Lambda Labs ($250 million) and Crusoe Energy ($100 million).
However, it is still significantly lower than major cloud providers like AWS, Microsoft Azure, and Google Cloud.
But CoreWeave has a certain competitive edge. The company operates on a rental model, where customers pay for the computing power they use per hour and also offers additional services like data storage, networking, and CPU compute.
The cost of a high-end H100 PCIe card for CoreWeave is roughly $30,000, which is rented out at an average of $4.25 per hour. Assuming an 80% utilisation rate, it would generate about $29,473 in annual revenue, roughly breaking even.
However, cheaper GPUs like the A40, purchased in bulk before the generative AI boom, could generate much higher margins. An A40 with a sticker price of $4,500 three years ago, rented out at $1.278 per hour, could generate $8,877 in annual revenue at 80% utilisation.
Together, a GPU reseller that rents GPUs from various sources and bundles them with AI training software, hit a $10M annual revenue run rate at the end of 2023, with 90% coming from their Forge product launched in June 2023.
Looking forward, CoreWeave’s advantage depends on the long-term state of the GPU industry and their ability to build a differentiated AI compute platform.
The current GPU shortage, driven by limitations at TSMC, is expected to last until March 2026, with a new $2.9B packaging facility operational in 2027.
CoreWeave’s infrastructure, designed from the ground up for GPU compute at scale, has shown promising results. Its HGX H100 instances delivered benchmark results 29x faster than the next-fastest competitor.
As AI workloads increase in size and complexity, CoreWeave’s specialisation in serving customers with the biggest compute needs could protect their moat even as GPU shortages ease.
Overcoming Bottlenecks
CoreWeave faces challenges, with a significant portion of its revenue coming from a small number of large customers like Microsoft and Inflection, while competitor Lambda Labs has a more extensive customer base.
The company also reduced its 2023 revenue projection, possibly indicating difficulties in securing sufficient chips or data centre space.
Bernstein analyst Stacy Rasgon has also noted that despite NVIDIA’s prioritisation of CoreWeave, the high demand for AI chips from various parties could lead to growing pains for cloud providers, who are struggling to keep pace with the AI boom.
Data centre operators face challenges in establishing facilities to accommodate NVIDIA’s power-hungry chips and meet the surging demand.
Despite these challenges, CoreWeave has been proactive in securing over 300 megawatts of data centre capacity and designing purpose-built facilities to handle AI workloads.
The company’s strong partnerships with NVIDIA and major customers, combined with its early-mover advantage, position it well for continued growth in the AI infrastructure market.
The post How This Crypto-Miner Turned AI Hyperscaler appeared first on AIM.
Looking at the dire need to build AI in India, by India, and for India, GreyOrange AI research scientist Guneet Singh Kohli went on a unique journey. He began working on Hugging Face’s Data is Better Together initiative in partnership with Daniel van Strien from Hugging Face and as the first step introduced Sanskriti Bench.
The aim of Sanskriti Bench is to develop an Indian cultural benchmark to test the increase of Indic AI models. By crafting a benchmark with the help of native speakers from different regions across India, the initiative aims to take into account the country’s cultural diversity.
The initiative is also being built with the help of Silo AI’s Dr Shantipriya Parida, who also created Odia Llama, Anindyadeep Sannigrahi from Prem AI, and Dr Kalyanamalini S, who is the language expert from Odia Generative AI.
Talking about the project with AIM, Kohli said that the most important and unique part about this project is that all of the data is novel, which means that it’s created by Indians from across the country to ensure diversity, accuracy, and quality of data. He said that this is not available in other datasets in Indic languages, which are essentially translations taken from English.
Apart from this, Kohli recently also partnered with GitHub and Save the Children to build AI tools for child safety and is preparing an AI system that can catch people who attempt to groom children online. “I write research papers, but eventually, there is no use if you can’t implement it for the people,” said Kohli.
The eventual goal Kohli has with this is to set up a global AI for Child Safety Lab, ahead of which he hopes to collaborate with several psychologists. “In India, children are using a lot of social media, and it becomes important for the country to also start talking about these,” he said, highlighting the importance of more conversations in the country as these are lacking when compared to the US and Europe.
The First Phase is Going Strong
According to the roadmap that Kohli laid out, the project is in its first phase. Currently, he is creating questions to build a dataset for benchmarking LLMs, which will then be hosted on the Hugging Face leaderboard.
In order to create the perfect questions, Kohli has taken the help of friends from different parts of the country. He gave an example of one friend from Bihar who provided questions in his native language, Maithili, along with the answers. However, the problem he highlighted was that these LLMs had a very big problem understanding context.
“We asked a question to an AI model about a festival from Bihar for which it was able to answer correctly with all the historical accuracy and the reasons for celebrating it. But when asked about more context on the festival, the model related the whole festival to Odisha,” said Kohli.
He explained that even though we can correct this by using our own knowledge of the culture, what about researchers who are using LLMs for research of Indian culture? “They would get it completely wrong,” said Kohli. Similarly, he highlighted how different states contribute to the country, like how Punjab is famous for agriculture and Gujarat is famous for driving the economy. All of this needs to be represented in the AI models as well with proper attribution.
To ensure these LLMs have geographical, cultural, historical, proverbial, and demographical knowledge about each part of the country in its native language, Kohli has started preparing the dataset.
He is working with several volunteers from Kashmir, Punjab, Kerala, and Assam to integrate the knowledge of each region into the dataset. “I am pushing for the idea that it needs to be completely human-driven,” he added, saying that he does not want to use synthetic data for creating questions as the foundation.
Currently, he is aiming for 500 questions per language and per region of the country, starting with 10 languages, which can be augmented using language models in later versions. “The beautiful part of India is each region has a unique language, which would make it diverse in itself,” said Kohli.
He is also working on incorporating figurative language, like the language used in proverbs, poems, similes, and other expressions, which are unique in each part of the country.
Indian Researchers Need to Get Together
With a BTech from Thapar Institute of Technology and working with a lot of non-profit organisations, Kohli’s motivation has always been to make technology work for the people. That is why he started working on the idea in 2020 with Cord.ai.
“I don’t want to call it my project. I eventually want to call it ‘by the people for the people’,” said Kohli, emphasising the fact that initiatives like these would be able to set up a benchmark created by Indians for anyone building AI models in the country, even if they are coming from outside the country.
“If anyone is creating an Indian model, it should be able to handle Indian culture,” said Kohli, highlighting that all the models coming up in the Indic AI space claiming to be the best need to be evaluated.
Talking about OpenAI coming to India and other Indian-based AI offerings such as Krutrim, Kohli said that it is important for researchers who are building different AI models in different languages to come together. He said that as the phases are completed, the initiative will be part of the community and everyone contributes to it.
Also, speaking about the recent launch of the Cohere Aya model in multiple languages, including Hindi, for which Kohli was one of the reviewers of the paper, he said, “If people from outside India can also do it for Indian languages, we being in India, why can’t we do it?”
The post Meet the Creator of Sanskriti Bench, Building Cultural AI for India with Hugging Face and GitHub appeared first on AIM.
Many of us start our programming journey with YouTube videos, and for the sake of simplicity, they often use print() statements to track bugs. That's fair enough, but as beginners adopt this habit, it can become problematic. Although these statements might work for simple scripts, as your codebase expands, this approach becomes highly inefficient. Therefore, in this article, I will introduce you to Python's built-in logging module, which solves this problem. We will see what logging is, how it differs from the print() statements, and we will also cover a practical example to fully understand its functionality.
Why Use the Logging Module Instead of Print()?
When we talk about debugging, the Python logging module provides much more detailed information than simple print() statements. This includes timestamps, module names, log levels, and line numbers where errors occurred, etc. These extra details help us understand the behavior of our code more effectively. The information we want to log depends on the needs of the application and the developer's preference. So, before we proceed further, let's discuss log levels and how to set them.
Logging Levels
You can control the amount of information you want to see using these log levels. Each log level has a numerical value that denotes its severity, with higher values indicating more severe events. For example, if you set your log level to WARNING, you're telling the logging module to only show you messages that are of WARNING level or higher. This means you won't see any DEBUG, INFO, or other less severe messages. This way, you can focus on the important events and ignore the noise
Here’s a table that shows the details of what each log level represents:
Log Level
Numerical Value
Purpose
DEBUG
10
Provides detailed information for diagnosing code-related issues, such as printing variable values and function call traces.
INFO
20
Used to confirm that the program is working as expected, like displaying startup messages and progress indicators.
WARNING
30
Indicates a potential problem that may not be critical to interrupt the program's execution but could cause issues later on.
ERROR
40
Represents an unexpected behavior of the code that impacts its functionality, such as exceptions, syntax errors, or out-of-memory errors.
CRITICAL
50
Denotes a severe error that can lead to the termination of the program, like system crashes or fatal errors.
Setting Up the Logging Module
To use the logging module, you need to follow some steps for configuration. This includes creating a logger, setting the logging level, creating a formatter, and defining one or more handlers. A handler basically decides where to send your log messages, such as to the console or a file. Let's start with a simple example. We're going to set up the logging module to do two things: first, it'll show messages on the console, giving us useful information (at the INFO level). Second, it'll save more detailed messages to a file (at the DEBUG level). I'd love it if you could follow along!
1. Setting the log level
The default level of the logger is set to WARNING. In our case, our two handlers are set to DEBUG and INFO levels. Hence, to ensure all messages are managed properly, we have to set the logger's level to the lowest level among all handlers, which, in this case, is DEBUG.
import logging # Create a logger logger = logging.getLogger(__name__) # Set logger level to DEBUG logger.setLevel(logging.DEBUG)
2. Creating a Formatter
You can personalize your log messages using formatters. These formatters decide how your log messages will look. Here, we will set up the formatter to include the timestamp, the log level, and the message content using the command below:
As discussed previously, handlers manage where your log messages will be sent. We will create two handlers: a console handler to log messages to the console and a file handler to write log messages to a file named 'app.log'.
Now that our setup is complete, let's test if it's working correctly before moving to the real-life example. We can log some messages as follows:
logger.debug('This is a debug message') logger.info('This is an info message') logger.warning('This is a warning message') logger.error('This is an error message') logger.critical('This is a critical message')
When you run this code, you should see the log messages printed to the console and written to a file named 'app.log', like this:
Console
2024-05-18 11:51:44,187 - INFO - This is an info message 2024-05-18 11:51:44,187 - WARNING - This is a warning message 2024-05-18 11:51:44,187 - ERROR - This is an error message 2024-05-18 11:51:44,187 - CRITICAL - This is a critical message
app.log
2024-05-18 11:51:44,187 - DEBUG - This is a debug message 2024-05-18 11:51:44,187 - INFO - This is an info message 2024-05-18 11:51:44,187 - WARNING - This is a warning message 2024-05-18 11:51:44,187 - ERROR - This is an error message 2024-05-18 11:51:44,187 - CRITICAL - This is a critical message
Logging User Activity in a Web Application
In this simple example, we will create a basic web application that logs user activity using Python's logging module. This application will have two endpoints: one for logging successful login attempts and the other to document failed ones (INFO for success and WARNING for failures).
1. Setting Up Your Environment
Before starting, set up your virtual environment and install Flask:
python -m venv myenv # For Mac source myenv/bin/activate #Install flask pip install flask
2. Creating a Simple Flask Application
When you send a POST request to the /login endpoint with a username and password parameter, the server will check if the credentials are valid. If they are, the logger records the event using logger.info() to signify a successful login attempt. However, if the credentials are invalid, the logger records the event as a failed login attempt using logger.error().
#Making Imports from flask import Flask, request import logging import os # Initialize the Flask app app = Flask(__name__) # Configure logging if not os.path.exists('logs'): os.makedirs('logs') log_file = 'logs/app.log' logging.basicConfig(filename=log_file, level=logging.DEBUG, format='%(asctime)s - %(levelname)s - %(message)s') log = logging.getLogger(__name__) # Define route and handler @app.route('/login', methods=['POST']) def login(): log.info('Received login request') username = request.form['username'] password = request.form['password'] if username == 'admin' and password == 'password': log.info('Login successful') return 'Welcome, admin!' else: log.error('Invalid credentials') return 'Invalid username or password', 401 if __name__ == '__main__': app.run(debug=True)
3. Testing the Application
To test the application, run the Python script and access the /login endpoint using a web browser or a tool like curl. For example:
Test Case 01
curl -X POST -d "username=admin&password=password" http://localhost:5000/login
Output
Welcome, admin!
Test Case 02
curl -X POST -d "username=admin&password=wrongpassword" http://localhost:5000/login
Output
Invalid username or password
app.log
2024-05-18 12:36:56,845 - INFO - Received login request 2024-05-18 12:36:56,846 - INFO - Login successful 2024-05-18 12:36:56,847 - INFO - 127.0.0.1 - - [18/May/2024 12:36:56] "POST /login HTTP/1.1" 200 - 2024-05-18 12:37:00,960 - INFO - Received login request 2024-05-18 12:37:00,960 - ERROR - Invalid credentials 2024-05-18 12:37:00,960 - INFO - 127.0.0.1 - - [18/May/2024 12:37:00] "POST /login HTTP/1.1" 200 -
Wrapping Up
And that wraps up this article. I strongly suggest making logging a part of your coding routine. It's a great way to keep your code clean and make debugging easier. If you want to dive deeper, you can explore the Python logging documentation for more features and advanced techniques. And if you're eager to enhance your Python skills further, feel free to check out some of my other articles:
Mastering Python: 7 Strategies for Writing Clear, Organized, and Efficient Code
8 Built-in Python Decorators to Write Elegant Code
Kanwal Mehreen Kanwal is a machine learning engineer and a technical writer with a profound passion for data science and the intersection of AI with medicine. She co-authored the ebook "Maximizing Productivity with ChatGPT". As a Google Generation Scholar 2022 for APAC, she champions diversity and academic excellence. She's also recognized as a Teradata Diversity in Tech Scholar, Mitacs Globalink Research Scholar, and Harvard WeCode Scholar. Kanwal is an ardent advocate for change, having founded FEMCodes to empower women in STEM fields.
More On This Topic
Using PyCaret’s New Time Series Module
Effective Testing for Machine Learning
Winning The Room: Creating and Delivering an Effective Data-Driven…
Prepare Your Data for Effective Tableau & Power BI Dashboards
The Fast and Effective Way to Audit ML for Fairness
Data Visualization Best Practices & Resources for Effective Communication
It was only a matter of time before Google injected AI into Chrome OS.
That time has now come.
Also: These top Chromebooks choices for students do it all
Chromebook Plus (a standard for higher-end Chrome OS hardware that includes exclusive features) is getting an update today that levels up the platform such that AI plays a more important role. The goal is to help you get the most out of AI when you're using Chrome OS for certain tasks.
Let's break down the most useful AI features coming to Chromebook Plus with this latest update.
1. Magic Editor in Google Photos
By now, you likely know what Magic Editor is and how it can help you create the perfect photo. Up until now, however, this feature was only available on phones. Get ready, because Magic Eraser is now available on Chromebook Plus.
Also: Magic Editor and other AI features in Google Photos are coming to your phone for free
You'll be able to select a photo in the Google Photos app, tap (or click) Magic Eraser, and start editing the image to your liking. You can reposition and resize objects, use contextual suggestions to improve the lighting and background, and completely reimagine your photos with a few clicks.
2. Gemini on your Chromebook
That's right, Gemini is now available on your Chromebook. Any time you need help with an idea, need to get an answer to a question, plan a trip, research a subject, and more, all you have to do is tap the Gemini icon on your app shelf and start interacting.
If you're a new Chromebook Plus user, you'll get the Google One AI Premium plan at no cost for 12 months. After that, you'll have to pay for the Gemini subscription. That plan includes access to Gemini Advance, 2TB of cloud storage, Gemini in Docs, Sheets, Slides, Gmail, and more.
3. Help Me Write
Help Me Write leverages Google's AI chops in all the places you write, such as websites, PDF forms, online applications, web apps, and more. When you need help writing, right-click (or two-finger tap) the text area to get suggestions or even get help changing the tone to fit your audience.
Also: Chrome now has a new AI writing tool to help you write almost anything online
Help Me Write helps generate text from scratch, using a prompt, or can help you rewrite existing text to make it more formal, shorten it, or totally rephrase it.
4. AI-generated wallpapers and video call backgrounds
With the help of AI, you'll be able to dream up just about any kind of image you want or need to serve as your Chromebook wallpaper or video call backgrounds.
You'll find some pre-built prompts included to help you build your backgrounds of all types (such as fun, whimsical, zen, and professional). Select what you want to see, and Google's AI will take it from there to generate an image specific to your prompt.
5. Quick access to Google Tasks
If you're a fan of Google Tasks, you might be happy to hear that you'll now have one-click access, via a built-in view of Google Tasks that makes it easy to add or check off todos.
Also: ChatGPT vs. Microsoft Copilot vs. Gemini: Which is the best AI chatbot?
Google Tasks will be accessible from the date icon on the bottom-right of your home screen, and will also be accessible across Google Workspace apps and devices. That means if you've added a task from Gmail on your Android phone, you can pick up where you left off on your Chromebook.
These new features will be available to Chromebook Plus devices on the latest Chrome OS version to be released on (or after — depending on your location) May 28, 2024.
Smartphone maker Nothing has appointed Vishal Bhola, former Whirlpool and Unilever executive, as the new president for its India operations.
Bhola brings over two decades of experience from his tenure at Unilever and Whirlpool. At Unilever, he held significant positions across various regions, including India, Southeast Asia, the USA, Africa, China, and London. Before joining Nothing, Bhola was the Managing Director for Whirlpool Corporation, overseeing profitability and growth in the Indian subcontinent.
I am thrilled to join as the President for @nothingindia. I look forward to working with @getpeid and the passionate team at Nothing. Let's make tech fun again!
— Vishal Bhola (@vishalbhola) May 28, 2024
Carl Pei, CEO and co-founder of Nothing, praised Bhola’s extensive background in the global consumer goods industry, calling him a valuable addition to the team. Pei, who also co-founded OnePlus, highlighted the company’s competitive position against giants like Apple, Samsung, Google, and OnePlus.
Founded in 2020, Nothing has sold over two million products worldwide, including its Phone (1). The company recently introduced its third smartphone, Phone (2a), which features industry-first ChatGPT integration when paired with any Nothing audio product. This new range of smartphones will be manufactured in India, aligning with the company’s long-term vision to establish India as its export hub.
Manu Sharma, who previously led Nothing’s India operations since February 2021, stepped down from his role as general manager and vice president in January. Sharma’s career includes extensive experience with Samsung Electronics and HP.
The post Vishal Bhola Appointed President of Nothing India appeared first on AIM.
