Latest news with #Github
Euronews
20 hours ago
- Euronews
This AI company can visualise your dreams. Here's how it works
A Dutch company says it has developed a way to use artificial intelligence (AI) to record dreams. Modem Works, an Amsterdam-based think tank and design studio, claims the Dream Recorder can capture dreams in 'ultra-low definition' and in any language. 'Wake up, speak your dream aloud … and watch it come to life in a dreamscape in the aesthetic of your choice,' the website for the project reads. How does it work? Modem Works says the project is 'Do-It-Yourself by Design.' It asks prospective users to download the open-source code, gather the hardware, 3D print the Dream Recorder's shell and assemble everything. Once assembled, users can double-tap to start a recording of themselves recalling their dream, and once finished, the dream will be generated. Another tap will play the generated dream and up to seven others that will be stored on a small 8-gigabyte processor. The company published the open source code on Github, a platform where coders share their projects, along with a list of the products they would need and where to buy them. The parts listed for the Dream Recorder include an HDMI screen, the 8-gigabyte processor, a micro SD card and a USB microphone. The approximate cost for all the parts to build a Dream Recorder is roughly €285, the developers wrote. The device would also require paying for the application programming interface (API) from OpenAI and AI video generation company LumaLabs to help generate the images for the dream. The developers estimate it would be less than $ 0.01 or $ 0.14 per dream, respectively, depending on the quality of the image. The Dream Recorder is the latest attempt to map out dreams with AI. In 2023, Japan's ATR Computational Neuroscience Laboratories developed an AI system that uses MRI scans to visualise and record dreams with a 60 per cent accuracy. Another pre-print study from the National University of Singapore and the Chinese University of Hong Kong in 2023 came to the same conclusion.
Miami Herald
16-06-2025
- Business
- Miami Herald
Analysts revamp forecast for Nvidia-backed AI stock
I have a virtual private server with several services running on it. It has replacements for Google Drive, Whatsapp, and Github (or Gitlab). Getting a sufficiently good internet connection that would allow me to use a real (on-premise) machine instead of a virtual one is very difficult where I live. I've been maintaining this server without any (serious) problems for a couple of years. However, in the past few months, the situation has changed, for the worse. Nothing brings me more joy than an occasional email from my VPS provider telling me that my server's CPU usage has been averaging at 98% for the last 2 hours. My server, which was almost invisible for a very long time, has become a target of scrapers and scanners. Related: Popular AI stock inks 5G network deal I am not alone in having this issue. Many prominent open-source projects had to protect themselves, too, and recently they started using "Anubis" for this. (Not the malware with the same name) Why the sudden change, you might ask? Well, an increasing number of companies think they will be the ones to create this 'incredible artificial intelligence.' So, they are scraping any website, regardless of whether its data is relevant and reliable. The more data they can collect, the better, seems to be the prevailing modus operandi. And once they're done collecting, throw everything into the blender and hope for the best. What if you are a little startup, with the aforementioned goals of writing incredible AI, and you've done the previous step of collecting the data, and now you just need that blender? Perhaps you have some investor money, but can't build that blender yourself. After all, Graphics cards used for AI training cost an arm and a leg. This is where CoreWeave (CRWV) comes in. Just like the VPS providers that enable people like me who can't use real machines for their servers to use their servers instead, CoreWeave enables companies that can't afford AI servers to do their AI training on its GPU mega clusters. More AI Stocks: Wall Street veteran doubles down on PalantirAnalysts double price target of new AI stock backed by NvidiaOpenAI teams up with legendary Apple exec Considering that the company's business model is "renting" Nvidia (NVDA) graphics cards, it is not surprising that the company has become Nvidia's largest holding, making up more than 78% of Nvidia's disclosed portfolio. Related: Micron makes massive $200 billion AI bet CoreWeave released its earnings report for Q1 2025 on May 14th. Here are the highlights: Revenue of $981.6 million, a 420% increase loss of $314.6 million, a 143% increase EBITDA of $606.1 million, a 480% increase YoY. Guidance for the full year 2025 was: Revenue from $4.9 billion to $5.1 billionCapital expenditures of $20 billion to $23 billion Bank of America analysts, Brad Sills and Carly Liu, shared their opinions on the CoreWeave stock. "In our view, the AI infrastructure [capital expenditures] growth rate is peaking, though still very healthy (estimates are likely to move higher on a larger base), led by OpenAI. OpenAI's ChatGPT is the single largest consumer of AI workloads and is growing at a rapid pace. Therefore, we see solid sustained demand in CoreWeave's AI infrastructure market," said analysts In Q1, CoreWeave expanded its deal with OpenAI bringing the total contract value to $15.9 billion. The company also signed a new hyperscaler customer in Q1. It has also increased the average contract duration to four and a half years from four years since 2024. Related: Apple WWDC underwhelms fans in a crucial upgrade Analysts forecasted $21 billion of negative free cash flow for the company through calendar year 2027, driven by high capital expenditures. CoreWeave funds the majority of its capital expenditures with debt. The company managed to lower the interest rate in the recent debt raise of $2 billion to 9.3%, from 11% in calendar year 2024. "However, this remains a small % of the total incremental debt required from here, raising some questions, in our view," continued analysts. Sills and Liu noted that the stock is trading at twenty-five times its calendar year 2027 EBIT estimate, which is a premium to the peer group that is trading at sixteen times the estimate. They set the new price objective for CoreWeave, raising their target from $76 to $185, which is 29 times their calendar year 2027 EBIT estimate (vs. 16x previously), or 0.4 times adjusted for 69% growth. That said, they cut their rating on the stock after CoreWeave's recent rally, arguing there's less room for shares to head higher. "We believe much of the near-term upside has been priced in and downgrade our rating to neutral from buy," concluded analysts. Related: Amazon's latest big bet may flop The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
NDTV
15-06-2025
- Business
- NDTV
Build A Portfolio Without A Full-Time Job To Attract Opportunities
Build Portfolio: Students aiming to land their first job can create an impressive portfolio by showcasing more than just work experience. Even without a full-time job, you can build a strong and appealing portfolio by consistently working on your skills and updating your work. Here's how you can craft a portfolio without holding a full-time position and attract opportunities. 1. Start With Personal Projects Work on personal projects related to your skill field and pick projects that could completely showcase your skills and be helpful in real world situations also. Share a detailed pdf on how your approached the problem and solved it. 2. Volunteer or Collaborate Offer your knowledge and skills to NGOs, college clubs, small businesses, or even friends. These unpaid collaborations still count as real work and can really help you strengthen your portfolio. For example, you can help a local cafe or a small food cart with their social media posts or offer to redesign your college fest website. 3. Take Part in Online Challenges or Hackathons Hackathons are recognized and conducted by various top companies like META and Microsoft. Securing the top spot or being in the top 10 performers in any hackathon helps recognize your skills around the world. You can showcase the hackathon challenge link in your portfolio. 4. Learn And Share Students can share their learning on online platforms like LinkedIn and Youtube. It helps you to improve with your skills and shows that you are passionate and continuously growing in the field. For example, writers can start a blog and developers can upload their projects on Github. 5. Include Case Studies For each project, briefly explain: The problem Your approach Tools/skills used Final result/outcome Building a strong and attractive portfolio takes time and commitment. Make sure you continue learning, upskilling, and regularly updating your portfolio with fresh work to stay relevant and job-ready.
Mint
10-06-2025
- Business
- Mint
Rahul Matthan: Digipin will displace workarounds to getting around
Whenever I order something online, I've learnt to provide as much location information as I can at checkout. As with most parts of India, the houses on my street are numbered somewhat at random (mine is 22/1 and my immediate neighbour's is 13), and there is no way anyone relying solely on GPS can find me without help. So they call, I explain, they get lost anyway, and we repeat this dance till somehow, with persistence and a lot of patience, the package finally reaches me. This is a story that repeats itself millions of times a day across the length and breadth of the country. Our addressing system is so broken that delivery agents have taken to calling even before they set out. Despite India being a top-five economy, we still navigate like medieval merchants, using directions like 'turn left at the paan shop." Also Read: Arguments against deploying DPI don't really hold up This is somewhat ironic considering that India has one of the world's most extensive postal networks. With 155,000 offices and 89% rural coverage, India Post can reach virtually every person in the country. But the challenge isn't reaching them, it is reaching them on time. In a world where next-day delivery is the norm and 10 minutes is fast becoming an expectation, our postal service is just not up to the task. So we've turned to logistics companies and their armies of delivery agents who zip around following turn-by-turn directions on mobile phones strapped to their handlebars to get parcels to us. And even if the miracles of modern technology get them close to us, the last mile often defeats them. Last week, the Department of Posts launched Digipin, a brand new digital addressing system that might be an answer to this problem. By dividing all of India into millions of tiny squares (about the size of a parking space), each with a unique 10-character name, all you would need to do is find the Digipin square that corresponds to the exact location of your doorstep and share the ten characters that define it with whoever needs to reach you. Also Read: DPI integration with AI will elevate the quality of public services as well as risks Unlike proprietary pinpoint systems like What3Words, Digipin is India's latest block of digital public infrastructure (DPI). Free, open and designed for national scale, it offers a foundation for addressing-as-a-service. Since its specifications are open source and easily accessible on Github, all it takes is a few API calls for applications to integrate with it. Which means that we no longer need to go through endless cycles of 'turn left at the temple, then right after the blue house," for deliveries to reach us. The real challenge will be adoption. Digipin names are unintelligible alphanumeric phrases like 'G4J-9K4-7LPY' that are hard to remember. If this is what we have to memorize to use it, we will get even more lost than we do today. What's needed is a way to translate these alphanumeric phrases into something we can recall. This is a problem that's already been solved. Internet websites are identified by 12-digit URLs that no one remembers. However, since all URLs are mapped to domain names that are easy to recall, we simply type those names into a browser for domain name servers (DNS) to translate into URLs that a computer can understand. Just as DNS unlocked the web for billions by translating forgettable URLs into names we could remember, a similar alias layer could unlock Digipin's mass adoption. Once we have this, I will be able to register a phrase like ' and map it to the precise Digipin coordinates of my front doorstep. Then all I need to do is share it with the delivery agent, who will then be able to navigate directly to my front door. Also Read: India's drive to globalize Digital Public Infrastructure: Time to take stock Digipin is designed to work with all Global Navigation Satellite Systems, but I am particularly pleased to see that it will work with India's own NavIC satellite constellation. In a 2017 Ex Machina article, I had pointed out that GPS is 'probably modern civilisation's single point of failure," given that it is a US Department of Defense system that underpins everything we rely on, from mobile networks to power grids. At this fractious time in global geopolitics, we should ensure that our digital location service relies on our own network of geo-positioning satellites. According to MIT Media Lab, India loses over 0.5% of its GDP annually because 80% of its addresses are described in relation to landmarks that lie anywhere between 50 and 1,500 metres away from their actual location. For an economy that wants to grow to $5 trillion by 2027, precision addressing isn't a luxury—it's a necessity. Also Read: Digital public infrastructure could help the world fight climate change As with every other DPI, the role of the government should be catalytic, not operational. Digipin's breakthrough moment will come when someone builds the addressing equivalent of Google Pay or PhonePe—with an interface so intuitive that using landmark-based directions feels as antiquated as paying with cash. Early adopters will, no doubt, be delivery-heavy businesses, but I'm far more keen to see the second-order effects: How precise addressing will reshape social coordination in ways that we have yet to imagine. Some 30 years ago, we couldn't imagine needing the internet; 20 years ago, mobile phones seemed like a luxury; 10 years ago, digital payments felt optional. Today, precise addressing feels like a nice-to-have, but tomorrow, we probably won't remember how we managed without it. For too long, we've been finding workarounds to getting around. Digipin is our chance to finally address our addressing challenge. The author is a partner at Trilegal and the author of 'The Third Way: India's Revolutionary Approach to Data Governance'. His X handle is @matthan.
Axios
20-05-2025
- Business
- Axios
AI agents will do the grunt work of coding
AI makers are flooding the market with a new wave of coding agents promising to relieve human programmers of busy work. The big picture: Automating the routine aspects of technical labor will almost certainly transform and downsize the tech industry workforce — but there's no guarantee it will alleviate software development's biggest headaches. Driving the news: Microsoft Monday announced a new AI coding agent for Github Copilot that's good for "time-consuming but boring tasks." "The agent excels at low-to-medium complexity tasks in well-tested codebases, from adding features and fixing bugs to extending tests, refactoring code, and improving documentation," Microsoft's post says. Github's move follows Friday's announcement by OpenAI of Codex, a "research preview" of a new coding agent that can "work on many tasks in parallel." Notably, the Github Copilot agent is powered not by Codex or any other tool from Microsoft partner OpenAI, but instead by Anthropic Claude 3.7 Sonnet, per Microsoft. The intrigue: Tech leaders have sent mixed messages on just how much work they see ahead for programmers. Amazon Web Services' then-boss Matt Garman caused a stir last year when he suggested the need for human coding could disappear within two years, However, he later told Axios that his comments were taken out of context. "I think it's incredibly exciting time for developers," he told us last year. "There's a whole bunch of work that developers do today that's not fun." "If you think about documenting your code, if you think about upgrading Java versions, if you think about looking for bugs, that's that's not what developers love doing. They love thinking about, 'How do I go solve problems?' " Why it matters: Business transformations that start in Silicon Valley usually make their way into the wider economy. Silicon Valley's "dogfooding" tradition ensures that it will avidly apply new technologies to its own business first. Both Microsoft and Google are now claiming that roughly 30% of the code they produce is AI-written. Coding agents, like other generative AI tools, continue to "hallucinate," or make stuff up. But programs, unlike other kinds of language products, have a built-in pass-fail test: Either they run or they don't. That gives programmers one early checkpoint to guard against bad code. Yes, but: AI-generated code likely also contains tons of other errors that don't show up today. That will cause nightmares in the future as programs age, get used more widely, or face unexpected tests from unpredictable users. Zoom out: The software industry's assumption that what works inside tech will work everywhere else could be sorely tested when these techniques get pushed out beyond Silicon Valley. AI's usefulness in writing code may not easily transfer to other kinds of work that are less abstract and more rooted in physical reality — witness the many setbacks and challenges the autonomous vehicle industry has faced. Between the lines: Nobody doubts that AI means tech firms will write more code using fewer employees. But no one yet knows exactly where these companies will continue to find competitive advantage. AI models are much more likely to be interchangeable than human organizations and cultures. What's next: As coding agents shoulder routine labor, product designers and creative engineers will use "vibe coding" — improvisational rough drafting via "throw it at the wall and see what works" AI prompting — to do fast prototyping of new ideas. The bottom line: The biggest challenges in creating software tend to arise from poorly conceived specifications and misinterpretations of data, both of which are often rooted in confusion over human needs.
