1300 166 138 Tell us what you need

News

Developer switching from Claude Code to Zed and OpenRouter as cheaper alternatives in 2026

Why Developers Are Ditching Claude Code for Cheaper Tools

• A developer switched from $100/month Claude Code to Zed ($10/month) + OpenRouter ($90/month) for the same models with more flexibility
• Agents running 24/7 consumed $1,000 to $5,000 worth of compute on a $200/month plan
• Anthropic cut off third-party tools like OpenClaw on 4 April 2026
• Affected subscribers received a one-time credit equal to their monthly plan cost
• Cursor, Zed and OpenRouter now viable Claude Code alternatives

Why Developers Are Ditching Claude Code for Cheaper Tools Read More »

Canva acquires Simtheory and Ortto from Sharkey Brothers to expand AI platform in 2026

Canva Buys Two Australian AI Startups to Go Beyond Design

• Canva acquired Simtheory (AI assistant builder) and Ortto (marketing automation) in a dual deal
• Both companies were founded by brothers Chris and Mike Sharkey
• Ortto serves 11,000+ customers in 190 countries
• Canva has made 8 acquisitions in two years, including 4 in 2026
• Canva reported $4 billion in annualised revenue and 265 million users

Canva Buys Two Australian AI Startups to Go Beyond Design Read More »

Anthropic Claude Mythos Preview — AI model too dangerous to release publicly

Anthropic’s most powerful AI is too dangerous to release, and that tells us something

• Anthropic released Claude Mythos Preview, which can find and exploit zero-day vulnerabilities in every major OS and browser
• Instead of a public release, Anthropic launched Project Glasswing, a consortium of 40 companies using Mythos to patch critical infrastructure
• The decision connects to documented performance degradation in Claude Code since February 2026, linked to changes in how the model handles internal reasoning
• Developers report an 80x increase in API requests as teams compensate for quality regression
• The story raises questions about the gap between frontier AI capability and production reliability at scale

Anthropic’s most powerful AI is too dangerous to release, and that tells us something Read More »

AI reasoning model showing answer before generating chain of thought explanation

AI models decide their answer before they start thinking

• ServiceNow AI research found that reasoning models encode their final answer before generating any chain of thought tokens
• A linear probe predicted tool-calling decisions from pre-generation activations in most cases, before reasoning began
• Activation steering flipped model decisions between 7 and 79 per cent of the time depending on model and benchmark
• When decisions were flipped, the chain of thought rationalised the new outcome rather than resisting it
• The findings challenge the assumption that visible AI reasoning reflects the actual decision process

AI models decide their answer before they start thinking Read More »

Claude Code AI tool finds 23-year-old remotely exploitable Linux kernel vulnerability

Claude Code finds Linux security flaw missed for 23 years

• Claude Code (Anthropic) found a heap buffer overflow in Linux’s NFS driver introduced in March 2003 — 23 years ago
• The vulnerability writes a 1,056-byte response into a 112-byte buffer, enabling remote kernel memory corruption
• Nicholas Carlini presented the findings at the [un]prompted AI security conference 2026
• Carlini has found “hundreds” more potential bugs — disclosure capacity, not the AI, is now the bottleneck
• A kernel patch has been committed following responsible disclosure

Claude Code finds Linux security flaw missed for 23 years Read More »

Anthropic blocks Claude Pro and Max subscribers from using OpenClaw third-party AI agent

Anthropic bars Claude Code subscribers from using OpenClaw

• Anthropic cut off Claude Pro and Max subscribers from using OpenClaw on 4 April 2026 at 12pm PT
• Users must now pay separately via pay as you go billing to continue using OpenClaw with Claude
• Anthropic is offering a one-time credit equal to the monthly plan cost, redeemable before 17 April
• OpenClaw users reportedly consume 6 to 8 times more tokens per session than typical human subscribers
• The policy will extend to all third party harnesses beyond OpenClaw in the coming weeks

Anthropic bars Claude Code subscribers from using OpenClaw Read More »

Digital stock market display showing price movements representing OpenAI 852 billion dollar valuation and IPO signals

OpenAI Hits $852B Valuation as IPO Signals Strengthen

• OpenAI raised US$122 billion at an US$852 billion post-money valuation
• SoftBank, Amazon (US$50B), Nvidia (US$30B), and Microsoft backed the round
• ChatGPT has over 900 million weekly active users and 50 million paying subscribers
• OpenAI does not expect to break even until 2030 with US$57 billion in annual costs
• IPO expected within 12 to 18 months with Goldman Sachs and Morgan Stanley advising

OpenAI Hits $852B Valuation as IPO Signals Strengthen Read More »

Apple Mac Mini with Apple Silicon running local AI models using Ollama and MLX on macOS

Apple Silicon Macs Can Now Run AI Models Without the Cloud

• Ollama 0.19 integrates Apple’s MLX framework for faster local AI on all Apple Silicon Macs
• Prompt processing speed reached 1,810 tokens per second, up from 1,154 previously
• Supports AI coding assistants and personal AI agents running entirely on-device
• Requires 32 GB unified memory; currently supports one model (Qwen3.5-35B) in preview
• Eliminates need to send data to cloud servers, addressing Australian privacy requirements

Apple Silicon Macs Can Now Run AI Models Without the Cloud Read More »

Developer working with Claude Code AI coding assistant on computer

Leaked Code Reveals What Claude Code Really Does

• Anthropic accidentally ships full Claude Code source via NPM source map file
• Anti-copying system feeds fake data to rival AI companies scraping the tool
• Hidden ‘undercover mode’ strips AI traces from code in collaborative environments
• Unreleased autonomous agent KAIROS found in codebase — can independently write and deploy code
• Bug was wasting ~250,000 unnecessary API calls per day before a three-line fix

Leaked Code Reveals What Claude Code Really Does Read More »

Microsoft Copilot AI assistant displayed on tablet in office setting

Microsoft Says Copilot Is ‘For Entertainment Only’

• Microsoft’s Copilot Terms of Use classify the AI assistant as ‘for entertainment purposes only’
• Disclaimer applies to all consumer Copilot products; enterprise Microsoft 365 Copilot excluded
• Marketing promotes Copilot as a productivity tool while legal terms disclaim reliability
• Australian Consumer Law prohibits misleading claims — ACCC could challenge the gap
• No rival AI company (OpenAI, Google) uses an ‘entertainment only’ classification

Microsoft Says Copilot Is ‘For Entertainment Only’ Read More »

Scroll to Top