ShieldFont is an open-source project that uses OpenType font features to replace words in HTML with grammatically equivalent but nonsense alternatives, making AI scrapers ingest poisoned data while human readers see normal text. It aims to deter unauthorized scraping by introducing uncertainty and cost, though it can be bypassed via OCR or targeted decryption and may impact SEO and accessibility.
ShieldFont leverages OpenType GSUB to substitute words with synonyms from the same grammatical category
HTML source appears garbled to scrapers, but users see the intended text
Perplexity expands its Model Council to the Computer platform, allowing users to select 2-8 AI models to independently analyze issues and produce a synthesis report. The feature, previously enterprise-only, is now available to Pro and Max subscribers with usage-based pricing.
Model Council lets users create a board of 2-8 AI models that independently tackle problems and produce a consensus/divergence report.
Expanded to the Computer platform, now accessible to Pro and Max users (not just enterprise).
A college professor tired of reading AI-generated essay responses hid a white-text instruction in an essay prompt, telling AI to insert 'Madagascar' nonsensically. All 35 students fell for it, submitting bizarre answers. The incident highlights concerns about AI's impact on critical thinking and academic integrity.
Professor embedded a hidden instruction for AI to insert 'Madagascar' nonsensically in the essay.
All 35 students submitted answers with odd uses of Madagascar.
A new study finds that leading AI models consistently land in the libertarian-left quadrant on the Political Compass, with Elon Musk's 'maximally truth-seeking' Grok showing a split personality. The likely cause: left-leaning training data.
Unslop tested 16 AI models on the Political Compass; all except Grok scored solidly in the libertarian-left quadrant.
Grok was the outlier: half its runs leaned right, half left, depending on the run.
IBM tells investors Q2 software weakness was temporary due to AI spending, with a third of deferred deals already closing. It launches Project Lightwell to fix legacy open source security using AI, at $1M annual subscription.
IBM claims software deal delays were due to customers prioritizing AI infrastructure, not abandonment.
CEO Krishna says a third of deferred deals closed in first three weeks of new quarter.
The European Commission announced two specification decisions requiring Google to open up search data to rivals and enhance Android AI interoperability for third-party assistants. Google objects, citing privacy and security concerns.
EU mandates Google to allow third-party AI assistants on Android to replace Gemini and perform actions on behalf of users.
Google must share anonymized search data with other search engines and AI chatbots to level the playing field.
Forrester warns that customers should brace for bigger software bills next year as software and AI vendors raise prices and pile on usage charges. The report highlights shifts to usage-based billing by Anthropic, OpenAI, GitHub, and Microsoft, and notes that despite layoffs, IT staffing costs remain high. It recommends adapting FinOps practices to manage unpredictable AI costs.
Forrester predicts software budgets will rise as AI vendors pass infrastructure costs via price hikes and usage fees.
Anthropic, OpenAI, GitHub, and Microsoft have moved to usage-based billing for some services.
Autopilot-Jobhunt is an AI tool that scans job postings, matches them to a user's resume, and notifies via Telegram. It can also generate tailored resumes and cover letters, but never auto-applies. It uses free AI models and prioritizes privacy.
Autopilot-Jobhunt automatically searches job listings and matches them to user profiles.
Users receive Telegram notifications with ranked job matches.
Startup MorphMind launches Academic Humanizer, a tool that uses AI to remove AI-writing tells from academic papers and grant proposals, raising concerns about integrity and the proliferation of low-quality AI-generated research.
Academic Humanizer is a Claude-based skill that rewrites AI-assisted drafts to sound more natural and personal.
It can be customized to match the user's previous writing style by referencing prior work.
Bengaluru-based developer Ashutosh Rath launches Atrophy, a CLI tool that treats coding skills like chess Elo ratings, offering drills in five areas to help developers measure and counteract the erosion of coding abilities due to AI reliance.
Atrophy CLI offers five skill drills: syntax recall, debugging, code reading, API memory, and decomposition, supporting Python and JavaScript with three difficulty levels.
Users take a 25-minute baseline test to get initial ratings (1200 each) and are advised to do 5-10 minute drills 2-3 times per week.
The open-source Godot engine team is rewriting its contribution policy to ban nearly all AI-generated contributions, citing demoralizing and unmanageable pull requests. New contributors must obtain explicit permission for significant changes, and AI agents are barred from communication channels.
Godot maintainers update policy to curb AI-generated pull requests
New contributors (fewer than 3 merged PRs) need maintainer approval for features or major refactoring
Microsoft introduces a lobby-based bot vetting system for Teams, requiring human approval for bots to join meetings, addressing security and privacy concerns. A registration path for legitimate ISV bots is also planned.
Microsoft adds a 'bouncer' to Teams meetings, requiring humans to verify bot identities in the lobby.
Unwanted bot attendance in sensitive meetings poses security and privacy risks.
Anthropic is killing off its existing Claude in Slack app in favor of an always-on agentic alternative dubbed Claude Tag that will listen and learn from everything it’s given access to. It can join Slack instances as a team member, access channels, tools, data, and codebases, and perform tasks via @Claude tags. It persists and learns context across channels. The old connector sunsets August 3; Enterprise and Teams customers get access today with migration credits.
Anthropic replaces Claude Slack app with Claude Tag, an always-on agentic AI that joins as a team member.
Claude Tag accesses channels, tools, data, and codebases, and can be delegated tasks via @Claude.
After a week-long test, the new Siri with Apple Intelligence degrades Spotlight search, requiring multiple taps for a simple web search, reminiscent of Google AI Overviews.
New Siri enables conversational interactions but ruins Spotlight by burying web search behind multiple taps.
Chat history automatically saved with no easy way to manage clutter.
Anthropic has beaten OpenAI to the IPO punch, filing a confidential S-1 with the SEC just after a $65B funding round pushed its valuation to $965B, surpassing OpenAI's $852B. The company's revenue leads despite smaller user base, but profitability remains uncertain. The IPO will offer the first real look at AI startup finances.
Anthropic filed for IPO confidentially after $65B series H, valuation $965B, surpassing OpenAI's $852B.
Claude Code drives revenue, but profit claims are questionable.
At Google I/O 2026, Google announced deeper AI integration into search, including expanded AI Overviews, AI Mode, and AI-powered ads. Critics fear this will bury traditional search results, harm the open web, and drive users to ad-blockers or alternative search engines.
Google announced expanded AI Overviews and a new AI Mode at I/O, pushing AI into more search queries.
New ad formats include inline AI ads and conversational discovery ads.
A provincial audit of 20 AI Scribe systems approved for Ontario healthcare providers found that 60% of evaluated systems mixed up prescribed drugs, 12 inserted incorrect drug info, and 9 fabricated information not discussed in recordings. Despite these errors, evaluation criteria weighted accuracy only 4% of total score, while domestic presence got 30%.
60% of AI Scribe systems mixed up prescribed drugs in patient notes.
9 out of 20 systems fabricated information not discussed in doctor-patient recordings.
A provincial audit of 20 AI Scribe systems approved for Ontario healthcare providers found that 60% of evaluated systems mixed up prescribed drugs, 12 inserted incorrect drug info, and 9 fabricated information not discussed in recordings. Despite these errors, evaluation criteria weighted accuracy only 4% of total score, while domestic presence got 30%.
60% of AI Scribe systems mixed up prescribed drugs in patient notes.
9 out of 20 systems fabricated information not discussed in doctor-patient recordings.
OpenAI president Greg Brockman testified in court that the company expects to burn $50 billion on computing power this year, using investor money. The AI firm remains unprofitable nearly four years after ChatGPT's launch, and has struggled to meet revenue targets. CEO Sam Altman has secured multi-billion dollar commitments from tech giants, but many are contingent on OpenAI leasing their compute infrastructure.
OpenAI expects to burn $50B on compute this year, funded by investors
Company has not turned a profit and misses revenue targets
Astera Labs unveiled a PCIe 6.0 switch for AI scale-up fabrics, offering a vendor-agnostic alternative to Nvidia's NVSwitch. The Scorpio X chip features 320 lanes and 5.12 TB/s bidirectional bandwidth, with in-network computing capabilities including Hypercast for MoE inference. Sampling now, production in H2 2026.
Astera Labs launches Scorpio X, a PCIe 6.0 switch with 320 lanes and 5.12 TB/s bandwidth.
It serves as a vendor-agnostic alternative to Nvidia's NVSwitch for AI scale-up fabrics.
Anthropic released financial agent templates for Claude, including skills, connectors, and subagents, to assist with finance tasks such as KYC screening, earnings review, and reconciliation. Despite leading benchmark scores, the company emphasizes the need for human oversight.
Anthropic's finance agent templates package skills (workflows), connectors (data access), and subagents (specialized Claude calls).
Templates cover ten tasks: pitch builder, meeting preparer, earnings reviewer, model builder, market researcher, valuation reviewer, general ledger reconciler, month-end closer, statement auditor, and KYC screener.
IBM updates Db2 Genius Hub with Google Vertex AI and Intel Gaudi to automate database management, promising cost reduction and less manual intervention. AI agents can propose and execute tasks with user approval, keeping human oversight.
IBM integrates Google Vertex AI and Intel Gaudi into Db2 Genius Hub for AI-driven automation.
AI agents can propose and execute database operations with user approval, aiming to cut management costs by 25% and manual intervention by 30%.