Tech News
← Home  ·  All topics

Ai

5527 GoKawiil briefs on this topic · RSS for ai

Pocketty launches iOS SSH terminal with alerts for stalled AI agents

Pocketty is a new iPhone and iPad app providing SSH terminal access designed for monitoring coding agents, built to work with a tool called herdr. The app sends notifications readable only on the user's phone when an agent is blocked or has finished a task, letting the user respond directly in a mobile terminal session. It offers a 14-day free trial followed by a one-time launch price of $99, discounted from $129.

OpenAI's new math proofs fall short of mathematicians' own guidelines

OpenAI released hundreds of claimed proofs to difficult open math problems using its proprietary models, saying it consulted the Advisory Group on Mathematics and Artificial Intelligence (AGMAI) to avoid prior controversies. But the release violated several of the group's own recommendations, including continuing to test proprietary models on hard problems, releasing chain-of-thought reasoning for only 10 of 719 manuscripts, and leaving 42% of proofs unformalized for human verification.

Hinton proposes FDA-style approval process for new AI models

Geoffrey Hinton, the computer scientist known as the 'Godfather of AI,' said on the Smart Girl Dumb Questions podcast that AI companies should be required to prove their products are safe to a federal regulator before public release, similar to drug approval. He noted that drug makers must spend roughly $1 billion proving safety to the FDA, and argued AI should face at least that level of scrutiny. Currently, AI firms rely on their own internal safety benchmarks rather than external review.

Entrepreneur op-ed: Better prompts, not pricier AI tools, drive better results

An Entrepreneur contributor argues that the gap between founders who get useful AI output and those who get 'AI slop' comes down to prompting technique rather than which AI subscription or tool they use. The piece, citing IBM research on prompt structure, says effective prompts need context, audience details, examples and a clearly defined task, and should be refined iteratively across multiple rounds.

Anthropic rolls out free OSS Scanner for automated open-source security checks

Anthropic has introduced OSS Scanner, a free opt-in service that runs periodic security scans on open-source projects using its most capable AI models, including Claude Mythos. The company says the reports are fully model-generated with no human review, which speeds up scanning but raises the risk of inaccurate or invalid findings.

Amazon replaces Fire tablets with new Alexa Tablet lineup starting at $230

Amazon is launching three new Alexa Tablets—the 8, 11, and 12 Pro—to replace its Fire tablet line, available in the US, Mexico and Canada from October 14, and in the UK, Germany, France, Italy and Spain five days later. Prices start at $230 for the 8.7-inch Alexa Tablet 8 and go up to $500 for the 12-inch Alexa Tablet 12 Pro, which Amazon says is its fastest tablet yet with improved processing and graphics. An anti-glare Nanomatte display option costs an extra $50, and a $90 stylus is available for the Pro model.

Edge Delta publishes open-source 'Project Arena' benchmark for AI SRE agents

Project Arena is a new vendor-neutral toolkit that spins up disposable Kubernetes test clusters, injects faults, and scores how well AI-driven site reliability tools investigate incidents. Its creators used the framework to compare Edge Delta's and Grafana's native AI investigation features, plus Claude running against each platform's observability CLI, across 21 Kubernetes incident scenarios, with all results judged by GPT-6-Astra. Edge Delta's native tool detected and investigated 18 of 21 scenarios versus Grafana's 12, with root-cause accuracy and mitigation-readiness also reported.

California enacts law limiting AI-only firing and discipline decisions

California Governor Gavin Newsom has signed SB 947, the 'No Robo Bosses Act,' which bars employers from letting automated decision systems alone fire or discipline workers. The law requires human verification of any AI-driven personnel decision and mandates employers disclose the data and reasoning behind it, including evaluations, work product and peer reviews. If a human reviewer finds the AI's output inaccurate or misleading, employers cannot act on it.

New 'Whistle' model brings on-device speech-to-text in 16.9 MB

A new speech recognition model called Whistle has been released for mobiles, wearables, robots, smart home devices, automotive systems and microcontrollers. It ships as a single 16.9 MB file that runs entirely on-device via CPU with no dependencies, handling transcription, word-level timestamps and speech embeddings for seven languages without sending audio off the device.

Jevman benchmark pits AI language models against classic Pac-Man ghosts

Jevman ran 100 games per AI model against scripted Pac-Man ghosts, with each game ending after three lost lives or a 5-minute cap. At every junction the model was asked to weigh the maze, pellets and ghost positions and output movement probabilities, with a 2-second deadline before a backup rule took over. Results were ranked by mean score with a 95% margin of error, and models within that margin were treated as tied.

StepFun's Step 5 Preview MoE model appears on OpenRouter with 1M context

StepFun has released Step 5 Preview, a sparse Mixture-of-Experts model with 27B active and 600B total parameters, now listed on OpenRouter. The model targets agentic workflows, supporting a 1-million-token context window for handling large codebases and documents. StepFun highlights its strength in software engineering, professional knowledge work, and finance-related tasks.

Factory AI engineer says he trained AI voice clone to read bedtime stories to newborn

Alvin Sng, a software engineer at startup Factory AI, posted on X that rather than reading a book aloud to his newborn himself, he offered to play an audiobook, then proposed training a voice model of himself to narrate it. He captioned the exchange 'Parenting hits different when you work in AI'; the post has since been deleted. The anecdote follows earlier comments from OpenAI's Sam Altman about using ChatGPT and AI-generated podcasts in place of everyday parent-child conversation.