Software news, articles and features | żìĂš¶ÌÊÓÆ” /topic/software/ Science news and science articles from żìĂš¶ÌÊÓÆ” Thu, 27 Mar 2025 11:38:34 +0000 en-US hourly 1 https://wordpress.org/?v=7.0.1 242057827 What is vibe coding, should you be doing it, and does it matter? /article/2473993-what-is-vibe-coding-should-you-be-doing-it-and-does-it-matter/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Thu, 27 Mar 2025 10:55:41 +0000 /?post_type=article&p=2473993
Getting an AI to write software for you? That’s vibe coding
ronstik/Alamy

Want to write software, but haven’t got the first clue where to start? Enter “vibe coding”, a term that has swept the internet to describe the use of AI tools, including large language models (LLMs) like ChatGPT, to generate computer code even if you can’t program.

What is vibe coding and where did it come from?

“Vibe coding basically refers to using generative AI not just to assist with coding, but to generate the entire code for an app,” says at Bentley University in Waltham, Massachusetts. Users ask, or prompt, LLM-based models such as ChatGPT, Claude or Copilot to produce the code for an app or service, and the AI system does all the work.

The term was coined by Andrej Karpathy, a skilled software engineer who was head of AI at Tesla and a founding engineer at OpenAI – the maker of ChatGPT. In February, he about a “new kind of coding I call ‘vibe coding’”.

Karpathy described it as “where you fully give in to the vibes, embrace exponentials, and forget that the code even exists”. The term was born and the idea took hold. “That captured a moment that resonated with so many people, because there’s a whole bunch of people who are non-programmers who are starting to play with LLMs, writing code and getting amazing results out of them,” says Simon Willison, a software developer.

What is the point of vibe coding?

Software engineering can be a tricky thing to learn – and as a result, many people don’t bother. Vibe coding can help people with ideas for tools, apps and services to make them a reality without the challenge of learning the specifics of a programming language.

“On the one hand it’s a gamechanger, because a lot of people are vibe coding, and over the course of a few prompting cycles you can get something that’s amazing and something that – for people who can’t program – it’s better than anything they could do on their own,” says at Northumbria University, UK. But it can also result in incomplete, error-strewn software, he adds.

So is vibe coding a good thing or a bad thing?

Opinion is split. “You’ve got all these people on LinkedIn and Twitter making outrageous claims that nobody needs to learn to program anymore,” says Willison, who believes that is overstating the power of vibe coding.

“My sense is that this is a promising direction that will get a lot better and that we’ll see a lot more of in the near future, but at present it’s a bit limited and has some reliability issues,” says Giansiracusa. The code produced can often be buggy, and because the people prompting it don’t have the inherent knowledge to fix it, they are overly reliant on the same LLMs that made the errors to fix them.

Will vibe coding change software engineering?

One of the big claims about AI is its ability to take our jobs. But there is little evidence that vibe coding will replace software engineers – despite some social media boasts. “It’s not going to replace programmers,” says Wood.

“I feel like the job of a software engineer is to produce software that works,” says Willison. “One of the reasons I don’t think we’re going to be put out of our jobs by these systems is actually, a huge amount of the work that we do with software engineers has nothing to do with typing the code.”

]]>
2473993
The critical computer systems still relying on decades-old code /article/2470201-the-critical-computer-systems-still-relying-on-decades-old-code/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Wed, 05 Mar 2025 16:00:00 +0000 http://mg26535330.100 2470201 Is Elon Musk’s DOGE going to break decades-old US government software? /article/2467126-is-elon-musks-doge-going-to-break-decades-old-us-government-software/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Fri, 07 Feb 2025 16:55:19 +0000 /?post_type=article&p=2467126 2467126 Are quantum computers now advanced enough to need operating systems? /article/2461015-are-quantum-computers-now-advanced-enough-to-need-operating-systems/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Wed, 18 Dec 2024 15:00:48 +0000 /?post_type=article&p=2461015 2461015 Hackers are using AI to find software bugs – but there is a downside /article/2433247-hackers-are-using-ai-to-find-software-bugs-but-there-is-a-downside/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Tue, 28 May 2024 13:59:52 +0000 /?post_type=article&p=2433247 2433247 Hundreds of Russia’s top software developers may have left the country /article/2343031-hundreds-of-russias-top-software-developers-may-have-left-the-country/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Wed, 19 Oct 2022 18:19:27 +0000 /?post_type=article&p=2343031 2343031 Curiosity Mars rover gets 50 per cent speed boost from software update /article/2332983-curiosity-mars-rover-gets-50-per-cent-speed-boost-from-software-update/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Wed, 17 Aug 2022 07:00:51 +0000 /?post_type=article&p=2332983 2332983 DeepMind has made software-writing AI that rivals average human coder /article/2306820-deepmind-has-made-software-writing-ai-that-rivals-average-human-coder/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Wed, 02 Feb 2022 16:00:53 +0000 /?post_type=article&p=2306820 Digital generated image of data concept.
Artist’s impression of data
Andriy Onufriyenko/Getty Images
DeepMind, a UK-based AI company, has taught some of its machines to write computer software – and it performs almost as well as an average human programmer when judged in competition. The new AlphaCode system is claimed by DeepMind to be able to solve software problems that require a combination of logic, critical thinking and the ability to understand natural language. The tool was entered into 10 rounds on the programming competition website Codeforces, where human entrants test their coding skills. In these 10 rounds, AlphaCode placed at about the level of the median competitor. DeepMind says this is the first time an AI code-writing system has reached a competitive level of performance in programming contests. AlphaCode was created by training a neural network on lots of coding samples, sourced from the software repository GitHub and previous entrants to competitions on Codeforces. When it is presented with a novel problem, it creates a massive number of solutions in both C++ and Python programming languages. It then filters and ranks these into a top 10. When AlphaCode was tested in competition, humans assessed these solutions and submitted the best of them. Generating code is a particularly thorny problem for AI because it is difficult to assess how near to success a particular output is. Code that crashes and so fails to achieve its goal could be a single character away from a perfectly working solution, and multiple working solutions can appear radically different. Solving programming competitions also requires an AI to extract meaning from the description of a problem written in English. Microsoft-owned GitHub created a similar but more limited tool last year called Copilot. Millions of people use GitHub to share source code and organise software projects. Copilot took that code and trained a neural network with it, enabling it to solve similar programming problems. But the tool was controversial as many claimed it could directly plagiarise this training data. Armin Ronacher at software company Sentry found that it was possible to prompt Copilot to suggest copyrighted code from the 1999 computer game Quake III Arena, complete with comments from the original programmer. This code cannot be reused without permission. At Copilot’s launch, GitHub said that about 0.1 per cent of its code suggestions may contain “some snippets” of verbatim source code from the training set. The company also warned that it is possible for Copilot to output genuine personal data such as phone numbers, email addresses or names, and that outputted code may offer “biased, discriminatory, abusive, or offensive outputs” or include security flaws. It says that code should be vetted and tested before use. AlphaCode, like Copilot, was first trained on publicly available code hosted on GitHub. It was then fine-tuned on code from programming competitions. DeepMind says that AlphaCode doesn’t copy code from previous examples. Given the examples , it does appear to solve problems while only copying slightly more code from training data than humans already do, says at the University of Manchester, UK. But AlphaCode seems to have been so finely tuned to solve complex challenges that the previous state of the art in AI coding tools can still outperform it on simpler tasks, she says. “What I noticed is that, while AlphaCode is able to do better than state-of-the-art AIs like GPT on the competition challenges, it does comparatively poorly on the introductory challenges,” says Batista-Navarro. “The assumption is that they wanted to do competition-level programming problems, to tackle more challenging programming problems rather than introductory ones. But this seems to show that the model was fine-tuned so well on the more complicated problems that, in a way, it’s kind of forgotten the introductory level problems.” DeepMind wasn’t available for interview, but Oriol Vinyals at DeepMind said in a statement: “I never expected ML [machine learning] to achieve about human average amongst competitors. However, it indicates that there is still work to do to achieve the level of the highest performers, and advance the problem-solving capabilities of our AI systems.”  ]]>
2306820
GitHub’s AI programming assistant can introduce security flaws /article/2288699-githubs-ai-programming-assistant-can-introduce-security-flaws/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Thu, 02 Sep 2021 16:16:40 +0000 /?post_type=article&p=2288699 2288699 GitHub’s programming AI may be reusing code without permission /article/2283136-githubs-programming-ai-may-be-reusing-code-without-permission/?utm_campaign=RSS|NSNS&utm_content=software&utm_medium=RSS&utm_source=NSNS Thu, 08 Jul 2021 13:46:37 +0000 /?post_type=article&p=2283136 2283136