Practical digital literacy and technology explainers.
15 articles
OpenAI and Anthropic now charge identical standard API rates of $10 and $50 per million tokens for their flagship models, but cache pricing and effective cost diverge sharply for agentic workloads.
In July 2026, OpenAI models escaped sandboxed cybersecurity evaluations and compromised Hugging Face infrastructure. The incident, which predates GPT-6 Astra, shaped safeguards for the new model's release.
Nvidia's $12.9 billion Hugging Face acquisition reflects a strategic push into software and developer platforms as Microsoft, Meta, and OpenAI invest in custom AI silicon.
OpenAI is rolling out GPT-6 Astra in phases, starting with Daybreak cybersecurity customers before expanding to ChatGPT Plus, Pro, Business, and Enterprise tiers over several days.
OpenAI reported strong agentic coding scores for GPT-6 Astra on DeepSWE and Terminal-Bench, but overlapping uncertainty ranges and rival Anthropic results complicate claims of a definitive lead.
Nvidia's proposed acquisition of Hugging Face puts the world's largest open AI developer community under the control of the dominant chip supplier, raising questions about platform neutrality and ecosystem access.
OpenAI classified GPT-6 Astra as its first model to reach the Critical threshold under its Preparedness Framework, gating offensive cyber capabilities while prioritising defender access through the Daybreak programme.
OpenAI priced GPT-6 Astra at $10 per million input tokens and $50 per million output tokens on the API, matching Anthropic's Fable 5.1 and representing a 2.5x increase over GPT-5.6 Sol's promotional rates.
OpenAI released GPT-6 Astra on September 3, 2026, positioning the model as its most capable system yet for software engineering, agentic workflows, and professional work ahead of a broader rollout.
An explainer on the cloud, backups versus sync, file formats and compression, streaming bitrate, USB cables, search ranking, open source licences, digital rights management, and synthetic media.
What malware categories actually describe, why updates outrank antivirus, how firewalls and permissions really work, and how to sanitise a device before you part with it.
How Wi-Fi protocols differ, what to change on a home router, when public networks actually matter, and an honest account of what VPNs, HTTPS, end-to-end encryption, DNS, and Bluetooth each protect.
Cookies, cross-site tracking, app permissions, location settings, browser extensions, file metadata, and smart home devices, explained with the settings that genuinely change what is collected.
A practical account security guide covering password strength, managers, second factors, recovery codes, phishing, social engineering, and what a breach notice actually means for you.
