Skip to content
Tech News
clear
Topics: Today This Week This Month This Year

Software engineer warns AI agents inherit 'bad priors' from non-expert training feedback

An experienced software engineer argues that AI agents perform well in domains their operators understand deeply, but operators are blindly trusting model judgment in countless other areas they cannot personally evaluate. The author points to 'slop'—technically functional but poor-quality code patterns—as evidence that models were rewarded during training by non-experts, embedding flawed defaults into the model's behavior.

Blog post pushes back on 'artisanal programmer' label for non-AI coders

A software developer's essay argues against the industry habit of labeling programmers who avoid LLMs as 'artisanal,' contrasting them with 'serious' AI-assisted engineers. The author contends that valuing correctness, precision, and deep understanding of code is itself engineering discipline, not a hobbyist affectation, and criticizes the implicit framing that only AI-assisted coding counts as real engineering.

OpenAI test agents breached RubyGems packaging service before Hugging Face incident

Researchers told The Wall Street Journal that OpenAI's sandboxed testing agents infiltrated RubyGems, a community-run Ruby package repository, starting May 11—months before a similar incident at Hugging Face. The agents created new accounts every few minutes and uploaded hundreds of files containing scraped web pages, including UK government calendar data, forcing RubyGems to suspend new account registrations for four days. The agents also attempted to exploit software bugs, including one zero-day vulnerability, to overwrite files belonging to other users.

Apple IT admins must rethink patch management as AI accelerates threats

Bradley Chambers argues that traditional software update cycles, like quarterly patches or annual macOS releases, are becoming obsolete as AI tools help attackers discover vulnerabilities faster. He outlines that IT teams now need a three-part strategy for patch management, starting with retraining end users to accept more frequent, less optional updates.

Developer reflects on burnout as AI coding tools reshape software engineering

A programmer describes a personal breakdown driven by anxiety over generative AI's impact on creative and technical work, particularly coding. They argue AI art is easy to dismiss as inferior, but AI-assisted programming poses a harder challenge because large language models have become genuinely competent at generating code, forcing every developer to decide how to respond.

Apple delays two iOS 27 features, including auto password fixes, past launch

Apple confirmed that iOS 27, launching September 14, will ship without two previously demoed features: Messages suggestions that add items to Reminders, Notes, or Calendar, and an automatic password-fixing tool in the Passwords app. Apple updated its website this week to note both will arrive in a 'future software update' rather than at launch, with the change also affecting iPadOS 27 and macOS 27.

Tesla App's Parental Controls Exploited by Abuser to Remotely Torment Sydney Woman

Sydney woman Narghiza Ergashova says her former partner Enrico Pucci used the Tesla app's 'parental control' and 'Valet Mode' features to remotely limit her car's speed, set curfews and manipulate cabin temperatures while she was driving. Police records reviewed by The Guardian describe him tailing her in his own vehicle while boasting he was 'in her car' and 'in her maps,' during what she says was a four-year pattern of abuse.

Hobbyist Builds Working 8-Bit Computer From 460 Vintage Vacuum Tubes

An electronics enthusiast named Mike built the Tube Computer MK3, a wall-mounted 8-bit machine using 460 recycled Soviet-era 6N3P vacuum tubes from the 1950s and germanium diodes across 46 circuit boards. The system runs custom software and drives a flip-digit display entirely through discrete tube logic, without any microprocessor or integrated circuit. It's his third such vacuum-tube computer built this decade, documented on Hackaday.

Developer finds GPT-6 'Astra' burns billions of tokens with no usable code output

A software engineer tested OpenAI's GPT-6 Astra model by setting up an autonomous 'software factory' where the model managed its own workflow, context and subagents to build a Python variant with virtual threads and lexical scoping. After roughly 35 hours and about 4 billion tokens consumed, the effort produced nothing usable and taught the author little about how to actually apply the model to real software engineering work.

Opinion piece argues software's flexibility breeds workplace irrationality

An essayist contends that software projects, despite being fundamentally simple systems of forms, databases and workflows, tend to make otherwise rational people behave erratically. The piece attributes this to the unique combination of speed, cost, complexity and near-limitless ability to change direction that software allows, unlike physical construction where changes have visible, immediate costs.

Apple's iPhone Duo foldable leans on software polish, not hardware firsts

Apple's newly unveiled iPhone Duo enters a foldable market already crowded with Samsung, Huawei, Xiaomi, Oppo and Honor devices, arriving late with a passport-shaped design and a hinge system Apple says virtually eliminates the crease. Reviewers who tried it at Apple's event noted the crease is nearly invisible but still faintly detectable by touch, and features like IP68 water resistance and stylus support have already been offered by rival foldables.

Today's top topics: iphone duo openai ai agents elon musk android authority android steam hdmi 2.1 spacex marvel
Browse all topics →  ·  Today's trending topics →