An anonymous reader quotes a report from Ars Technica: A Polish programmer running on fumes recently accomplished what may soon become impossible: beating an advanced AI model from OpenAI in a head-to-head coding competition. The 10-hour marathon left him "completely exhausted." On Wednesday, programmer Przemysaw Debiak (known as "Psyho"), a former OpenAI employee, narrowly defeated the custom AI model in the AtCoder World Tour Finals 2025 Heuristic contest in Tokyo. AtCoder, a Japanese platform that hosts competitive programming contests and maintains global rankings, held what may be the first contest where an AI model competed directly against top human programmers in a major onsite world championship. During the event, the maker of ChatGPT participated as a sponsor and entered an AI model in a special exhibition match titled "Humans vs AI." Despite the tireless nature of silicon, the company walked away with second place.

The competition required contestants to solve a single complex optimization problem over 600 minutes. The contest echoes the American folk tale of John Henry, the steel-driving man who raced against a steam-powered drilling machine in the 1870s. Like Henry's legendary battle against industrial automation, Debiak's victory represents a human expert pushing themselves to their physical limits to prove that human skill still matters in an age of advancing AI. Both stories feature exhausting endurance contests -- Henry drove steel spikes for hours until his heart gave out, while Debiak coded for 10 hours on minimal sleep. The parallel extends to the bittersweet nature of both victories: Henry won his race but died from the effort, symbolizing the inevitable march of automation, while Debiak's acknowledgment that humanity prevailed "for now" suggests he recognizes this may be a temporary triumph against increasingly capable machines. While Debiak won 500,000 yen and survived his ordeal better than the legendary steel driver, the AtCoder World Tour Finals pushes humans and AI models to their limits through complex optimization challenges that have no perfect solution -- only incrementally better ones. "Humanity has prevailed (for now!)," wrote Debiak on X, noting he had little sleep while competing in several competitions across three days. "I'm completely exhausted. ... I'm barely alive."

Robinhood CEO Vlad Tenev has said that the majority of his company's new code is written by AI, with "close to 100%" of engineers using AI code editors. Speaking on the 20VC podcast, Tenev estimated around 50% of new code at the trading platform is AI-generated.

Tenev said the 50% figure is imprecise due to advanced "agentic" code editors that have made it difficult to distinguish human-written from AI-generated code. The company has progressed from GitHub Copilot to Cursor and now Windsurf, where "nearly all of the code is written by AI," he said. Tenev estimated only a "minority" of new code at Robinhood is written by humans.

A survey of 500 IT asset managers in organizations that use Oracle Java has found that 73% have been audited in the last three years. From a report: At the same time, nearly eight out of 10 Oracle Java users said they had migrated, or planned to shift, to open source Java to try to avoid the risk and high costs of the dominant vendor's development and runtime environments.

Oracle introduced a paid subscription for Java in September 2018, and in January 2023, it decided to switch its pricing model to per employee rather than per user, creating a steep price hike for many users. In July 2023, Gartner recorded users experiencing price increases of between two and five times when they switched to the new licensing model.

Two years later, the survey conducted by market research firm Dimensional Research showed only 14% of Oracle Java users intended to stick with the vendor's subscription model.

Chinese AI startup Moonshot AI has released Kimi K2, a trillion-parameter open-source language model that outperforms GPT-4 in key benchmarks with particularly strong performance on coding and autonomous agent tasks. VentureBeat reports: The new model, called Kimi K2, features 1 trillion total parameters with 32 billion activated parameters in a mixture-of-experts architecture. The company is releasing two versions: a foundation model for researchers and developers, and an instruction-tuned variant optimized for chat and autonomous agent applications. "Kimi K2 does not just answer; it acts," the company stated in its announcement blog. "With Kimi K2, advanced agentic intelligence is more open and accessible than ever. We can't wait to see what you build."

The model's standout feature is its optimization for "agentic" capabilities -- the ability to autonomously use tools, write and execute code, and complete complex multi-step tasks without human intervention. In benchmark tests, Kimi K2 achieved 65.8% accuracy on SWE-bench Verified, a challenging software engineering benchmark, outperforming most open-source alternatives and matching some proprietary models. [...] On LiveCodeBench, arguably the most realistic coding benchmark available, Kimi K2 achieved 53.7% accuracy, decisively beating DeepSeek-V3's 46.9% and GPT-4.1's 44.7%. More striking still: it scored 97.4% on MATH-500 compared to GPT-4.1's 92.4%, suggesting Moonshot has cracked something fundamental about mathematical reasoning that has eluded larger, better-funded competitors.

But here's what the benchmarks don't capture: Moonshot is achieving these results with a model that costs a fraction of what incumbents spend on training and inference. While OpenAI burns through hundreds of millions on compute for incremental improvements, Moonshot appears to have found a more efficient path to the same destination. It's a classic innovator's dilemma playing out in real time -- the scrappy outsider isn't just matching the incumbent's performance, they're doing it better, faster, and cheaper.

An anonymous reader shared this report from InfoWorld: Tiobe CEO Paul Jansen says Ada, a system programming language whose initial development dates back to the late 1970s, could outlast similarly aged languages like Visual Basic, Perl, and Fortran in the language popularity race.

In comments on this month's Tiobe language popularity index, posted July 9, Jansen said the index has not seen much change among leading languages such as Python, C#, and Java over the past two years. But there is more movement among older languages such as Visual Basic, SQL, Fortran, Ada, Perl, and Delphi, said Jansen. Every time one of these languages is expected to stay in the top 10, it is replaced by another language, he said. Even more remarkably, newer languages have yet to rise above them. "Where are Rust, Kotlin, Dart, and Julia? Apparently, established languages are hot."
"Which one will win? Honestly, this is very hard to tell," Jansen writes, "but I would put my bets on Ada. With the ever-stronger demands on security, Ada is, as a system programming language in the safety-critical domain, likely the best survivor."

Perhaps proving his point, one year ago, Ada was ranked #24 — but on this month's index it ranks #9. (Whereas the eight languages above it all remain in the exact same positions they held a year ago...)

1. Python
2. C++
3. C
4. Java
5. C#
6. JavaScript
7. Go
8. Visual Basic
9. Ada
10. Delphi/Object Pascal

An anonymous reader quotes a report from Reuters: Contrary to popular belief, using cutting-edge artificial intelligence tools slowed down experienced software developers when they were working in codebases familiar to them, rather than supercharging their work, a new study found. AI research nonprofit METR conducted the in-depth study on a group of seasoned developers earlier this year while they used Cursor, a popular AI coding assistant, to help them complete tasks in open-source projects they were familiar with. Before the study, the open-source developers believed using AI would speed them up, estimating it would decrease task completion time by 24%. Even after completing the tasks with AI, the developers believed that they had decreased task times by 20%. But the study found that using AI did the opposite: it increased task completion time by 19%. The study's lead authors, Joel Becker and Nate Rush, said they were shocked by the results: prior to the study, Rush had written down that he expected "a 2x speed up, somewhat obviously." [...]

The slowdown stemmed from developers needing to spend time going over and correcting what the AI models suggested. "When we watched the videos, we found that the AIs made some suggestions about their work, and the suggestions were often directionally correct, but not exactly what's needed," Becker said. The authors cautioned that they do not expect the slowdown to apply in other scenarios, such as for junior engineers or engineers working in codebases they aren't familiar with. Still, the majority of the study's participants, as well as the study's authors, continue to use Cursor today. The authors believe it is because AI makes the development experience easier, and in turn, more pleasant, akin to editing an essay instead of staring at a blank page. "Developers have goals other than completing the task as soon as possible," Becker said. "So they're going with this less effortful route."

The University of Washington's Paul G. Allen School of Computer Science & Engineering is overhauling its approach to computer science education as AI reshapes the tech industry. Director Magdalena Balazinska has declared that "coding, or the translation of a precise design into software instructions, is dead" because AI can now handle that work.

The Pacific Northwest's premier tech program now allows students to use GPT tools in assignments, requiring them to cite AI as a collaborator just as they would credit input from a fellow student. The school is considering "coordinated changes to our curriculum" after encouraging professors to experiment with AI integration.

BrianFagioli shares a report from NERDS.xyz: Android is changing how it gives developers access to early features. The company is replacing its old Developer Preview model with a new Canary channel that provides rolling updates all year long. This new approach is meant to give developers earlier and more consistent access to experimental tools and APIs.

Previously, Developer Previews had to be manually flashed onto devices. They only ran during the earliest stages of each release cycle and stopped once Android entered the beta phase. That meant promising features that were not quite ready for beta had nowhere to go and no way to collect feedback. The Canary channel solves that by running in parallel with the existing beta program and delivering over the air updates automatically.

BrianFagioli shares a report from NERDS.xyz: Red Hat has introduced a new option that gives developers a fast lane to enterprise-grade Linux without needing to go through IT. The new release, called Red Hat Enterprise Linux for Business Developers, is now available for free. It offers direct, self-serve access to the same operating system used in production environments, specifically for business-focused development and testing.

The offering is part of the Red Hat Developer Program and is designed to reduce friction between development and operations teams. Developers can now build and test applications on the same platform that powers critical systems across physical servers, virtual machines, cloud deployments, and edge devices. [...] Each registered user can deploy up to 25 instances, whether virtual, physical, or cloud-based. The program includes signed and curated developer content such as programming languages, open source tools, and databases. Red Hat also includes Podman Desktop, its go-to container development tool, allowing users to work with containers that can closely match production environments.

While access is free, developers can choose to purchase support plans that tap into Red Hat's Linux expertise. This could appeal to developers working in business units or teams that want to build quickly without waiting on formal IT approval. This new option complements Red Hat's existing free Developer Subscription for Individuals and the Enterprise Developer Subscription for Teams, which is available through Red Hat reps or partners.

An anonymous reader quotes a report from VentureBeat: Hugging Face, the $4.5 billion artificial intelligence platform that has become the GitHub of machine learning, announced Tuesday the launch of Reachy Mini, a $299 desktop robot designed to bring AI-powered robotics to millions of developers worldwide. The 11-inch humanoid companion represents the company's boldest move yet to democratize robotics development and challenge the industry's traditional closed-source, high-cost model.

The announcement comes as Hugging Face crosses a significant milestone of 10 million AI builders using its platform, with CEO Clement Delangue revealing in an exclusive interview that "more and more of them are building in relation to robotics." The compact robot, which can sit on any desk next to a laptop, addresses what Delangue calls a fundamental barrier in robotics development: accessibility. "One of the challenges with robotics is that you know you can't just build on your laptop. You need to have some sort of robotics partner to help in your building, and most people won't be able to buy $70,000 robots," Delangue explained, referring to traditional industrial robotics systems and even newer humanoid robots like Tesla's Optimus, which is expected to cost $20,000-$30,000.

Reachy Mini emerges from Hugging Face's April acquisition of French robotics startup Pollen Robotics, marking the company's most significant hardware expansion since its founding. The robot represents the first consumer product to integrate natively with the Hugging Face Hub, allowing developers to access thousands of pre-built AI models and share robotics applications through the platform's "Spaces" feature. [...] Reachy Mini packs sophisticated capabilities into its compact form factor. The robot features six degrees of freedom in its moving head, full body rotation, animated antennas, a wide-angle camera, multiple microphones, and a 5-watt speaker. The wireless version includes a Raspberry Pi 5 computer and battery, making it fully autonomous. The robot ships as a DIY kit and can be programmed in Python, with JavaScript and Scratch support planned. Pre-installed demonstration applications include face and hand tracking, smart companion features, and dancing moves. Developers can create and share new applications through Hugging Face's Spaces platform, potentially creating what Delangue envisions as "thousands, tens of thousands, millions of apps." Reachy Mini's $299 price point could significantly transform robotics education and research. "Universities, coding bootcamps, and individual learners could use the platform to explore robotics concepts without requiring expensive laboratory equipment," reports VentureBeat. "The open-source nature enables educational institutions to modify hardware and software to suit specific curricula. Students could progress from basic programming exercises to sophisticated AI applications using the same platform, potentially accelerating robotics education and workforce development."

"... For the first time, a major AI platform is betting that the future of robotics belongs not in corporate research labs, but in the hands of millions of individual developers armed with affordable, open-source tools."

"Apple quietly dropped a new AI model on Hugging Face with an interesting twist," writes 9to5Mac. "Instead of writing code like traditional LLMs generate text (left to right, top to bottom), it can also write out of order, and improve multiple chunks at once."

"The result is faster code generation, at a performance that rivals top open-source coding models." Traditionally, most LLMs have been autoregressive. This means that when you ask them something, they process your entire question, predict the first token of the answer, reprocess the entire question with the first token, predict the second token, and so on. This makes them generate text like most of us read: left to right, top to bottom... An alternative to autoregressive models is diffusion models, which have been more often used by image models like Stable Diffusion. In a nutshell, the model starts with a fuzzy, noisy image, and it iteratively removes the noise while keeping the user request in mind, steering it towards something that looks more and more like what the user requested...

Lately, some large language models have looked to the diffusion architecture to generate text, and the results have been pretty promising... This behavior is especially useful for programming, where global structure matters more than linear token prediction... [Apple] released an open-source model called DiffuCode-7B-cpGRPO, that builds on top of a paper called DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation, released just last month... [W]ith an extra training step called coupled-GRPO, it learned to generate higher-quality code with fewer passes. The result? Code that's faster to generate, globally coherent, and competitive with some of the best open-source programming models out there.

Even more interestingly, Apple's model is built on top of Qwen2.5-7B, an open-source foundation model from Alibaba. Alibaba first fine-tuned that model for better code generation (as Qwen2.5-Coder-7B), then Apple took it and made its own adjustments. They turned it into a new model with a diffusion-based decoder, as described in the DiffuCoder paper, and then adjusted it again to better follow instructions. Once that was done, they trained yet another version of it using more than 20,000 carefully picked coding examples.
"Although DiffuCoder did better than many diffusion-based coding models (and that was before the 4.4% bump from DiffuCoder-7B-cpGRPO), it still doesn't quite reach the level of GPT-4 or Gemini Diffusion..." the article points out.

But "the bigger point is this: little by little, Apple has been laying the groundwork for its generative AI efforts with some pretty interesting and novel ideas."

An anonymous reader shared this report from the San Francisco Standard: About an hour into my meeting with the undisputed hackathon king of San Francisco, Rene Turcios asked if I wanted to smoke a joint with him. I politely declined, but his offer hardly surprised me. Turcios has built a reputation as a cannabis-loving former professional Yu-Gi-Oh! player who resells Labubus out of his Tenderloin apartment when he's not busy attending nearly every hackathon happening in the city. Since 2023, Turcios, 29, has attended more than 200 events, where he's won cash, software credits, and clout. "I'm always hustling," he said.

The craziest part: he doesn't even know how to code.

"Rene is the original vibe coder," said RJ Moscardon, a friend and fellow hacker who watched Turcios win second place at his first-ever hackathon at the AGI House mansion in Hillsborough. "All the engineers with prestigious degrees scoffed at him at first. But now they're all doing exactly the same thing...." Turcios was vibe coding long before the technique had a name — and was looked down upon by longtime hackers for using AI. But as Tiger Woods once said, "Winning takes care of everything...."

Instead of vigorously coding until the deadline, he finished his projects hours early by getting AI to do the technical work for him. "I didn't write a single line of code," Turcios said of his first hackathon where he prompted ChatGPT using plain English to generate a program that can convert any song into a lo-fi version. When the organizers announced Turcios had won second place, he screamed in celebration.... "I realized that I could compete with people who have degrees and fancy jobs...."

Turcios is now known for being able to build anything quickly. Businesses reach out to him to contract out projects that would take software engineering teams weeks — and he delivers in hours. He's even started running workshops to teach non-technical groups and experienced software engineers how to get the most out of AI for coding.
"He grew up in Missouri to parents who worked in an international circus, taming bears and lions..."

"A computer science degree used to be a golden ticket to the promised land of jobs," a college senior tells the New York Times. But "That's no longer the case."

The article notes that in the last three years there's been a 65% drop from companies seeking workers with two years of experience or less (according to an analysis by technology research/education organization CompTIA), with tech companies "relying more on AI for some aspects of coding, eliminating some entry-level work."

So what do college professors teach when AI "is coming fastest and most forcefully to computer science"? Computer science programs at universities across the country are now scrambling to understand the implications of the technological transformation, grappling with what to keep teaching in the AI era. Ideas range from less emphasis on mastering programming languages to focusing on hybrid courses designed to inject computing into every profession, as educators ponder what the tech jobs of the future will look like in an AI economy... Some educators now believe the discipline could broaden to become more like a liberal arts degree, with a greater emphasis on critical thinking and communication skills.

The National Science Foundation is funding a program, Level Up AI, to bring together university and community college educators and researchers to move toward a shared vision of the essentials of AI education. The 18-month project, run by the Computing Research Association, a research and education nonprofit, in partnership with New Mexico State University, is organising conferences and roundtables and producing white papers to share resources and best practices. The NSF-backed initiative was created because of "a sense of urgency that we need a lot more computing students — and more people — who know about AI in the workforce," said Mary Lou Maher, a computer scientist and a director of the Computing Research Association.

The future of computer science education, Maher said, is likely to focus less on coding and more on computational thinking and AI literacy. Computational thinking involves breaking down problems into smaller tasks, developing step-by-step solutions and using data to reach evidence-based conclusions. AI literacy is an understanding — at varying depths for students at different levels — of how AI works, how to use it responsibly and how it is affecting society. Nurturing informed skepticism, she said, should be a goal.
The article raises other possibilities. Experts also suggest the possibility of "a burst of technology democratization as chatbot-style tools are used by people in fields from medicine to marketing to create their own programs, tailored for their industry, fed by industry-specific data sets." Stanford CS professor Alex Aiken even argues that "The growth in software engineering jobs may decline, but the total number of people involved in programming will increase."

Last year, Carnegie Mellon actually endorsed using AI for its introductory CS courses. The dean of the school's undergraduate programs believes that coursework "should include instruction in the traditional basics of computing and AI principles, followed by plenty of hands-on experience designing software using the new tools."

"Microsoft has released the source code for the GitHub Copilot Chat extension for VS Code under the MIT license," reports BleepingComputer. This provides the community access to the full implementation of the chat-based coding assistant, including the implementation of "agent mode," what contextual data is sent to large language models (LLMs), and the design of system prompts. The GitHub repository hosting the code also details telemetry collection mechanisms, addressing long-standing questions about data transparency in AI-assisted coding tools...

As the VS Code team explained previously, shifts in AI tooling landscape like the rapid growth of the open-source AI ecosystem and a more level playing field for all have reduced the need for secrecy around prompt engineering and UI design. At the same time, increased targeting of development tools by malicious actors has increased the need for crowdsourcing contributions to rapidly pinpoint problems and develop effective fixes. Essentially, openness is now considered superior from a security perspective.
"If you've been hesitant to adopt AI tools because you don't trust the black box behind them, this move opensources-github-copilot-chat-vscode/offers something rare these days: transparency," writes Slashdot reader BrianFagioli" Now that the extension is open source, developers can audit how agent mode actually works. You can also dig into how it manages your data, customize its behavior, or build entirely new tools on top of it. This could be especially useful in enterprise environments where compliance and control are non negotiable.

It is worth pointing out that the backend models powering Copilot remain closed source. So no, you won't be able to self host the whole experience or train your own Copilot. But everything running locally in VS Code is now fair game. Microsoft says it is planning to eventually merge inline code completions into the same open source package too, which would make Copilot Chat the new hub for both chat and suggestions.

Software engineer Sean Goedecke argues that AI coding agents have already been commoditized because they require no special technical advantages, just better base models. He writes: All of a sudden, it's the year of AI coding agents. Claude released Claude Code, OpenAI released their Codex agent, GitHub released its own autonomous coding agent, and so on. I've done my fair share of writing about whether AI coding agents will replace developers, and in the meantime how best to use them in your work. Instead, I want to make what I think is now a pretty firm observation: AI coding agents have no secret sauce.

[...] The reason everyone's doing agents now is the same reason everyone's doing reinforcement learning now -- from one day to the next, the models got good enough. Claude Sonnet 3.7 is the clear frontrunner here. It's not the smartest model (in my opinion), but it is the most agentic: it can stick with a task and make good decisions over time better than other models with more raw brainpower. But other AI labs have more agentic models now as well. There is no moat.

There's also no moat to the actual agent code. It turns out that "put the model in a loop with a 'read file' and 'write file' tool" is good enough to do basically anything you want. I don't know for sure that the closed-source options operate like this, but it's an educated guess. In other words, the agent hackers in 2023 were correct, and the only reason they couldn't build Claude Code then was that they were too early to get to use the really good models.