Business, Deals & Funding
Claude Code Changelog
v2.1.178
Version 2.1.178 of Claude Code introduces Tool(param:value) syntax for permission rules to match tool input parameters with wildcard support (e.g., Agent(model:opus) to block Opus subagents). Skills in nested .claude/skills directories now load when working on files in those directories, with name clashes resolved by prefixing with a colon to keep both available. Nested .claude/ directories now follow a closest-to-working-directory-wins rule for agent, workflow, and output-style configurations, with project-scope workflow saves targeting the closest existing directory.
Why it matters
This is a solid incremental update focused on improving the flexibility and granularity of Claude Code's configuration system. The Tool(param:value) syntax for permission rules is particularly useful for fine-grained access control, such as blocking specific model subagents. The nested .claude/ directory improvements show thoughtful handling of monorepo and complex project structures, with sensible conflict resolution strategies. These are the kinds of quality-of-life improvements that make the…
TechCrunch AI
Malaysia’s AI agent-powered messaging app Respond.io raises $62.5M, eyes acquisitions
Respond.io, a Malaysia-headquartered AI agent-powered customer conversation management platform, has raised a $62.5 million Series B round led by Camber Partners with participation from Endeavor Catalyst and existing investors. Founded in 2017 by Gerardo Salandra, Hassan Ahmed, and Iaroslav Kudritskiy, the company helps mid- to large-sized B2C businesses manage customer conversations across messaging channels like WhatsApp, Instagram, TikTok, and others. The platform uses AI agents to handle high volumes of customer inquiries, qualify leads, and close sales. Respond.io has grown to $35 million in ARR with 169% year-over-year growth and a 30% profit margin, processing 2 billion messages per quarter. Unlike competitors that charge per seat, Respond.io charges per conversation volume, making its model resilient to AI replacing human agents. The company plans to use the funding for hiring,…
Why it matters
Respond.io presents a compelling growth story with impressive metrics — 169% YoY growth, $35M ARR, and 30% profit margins are exceptional for a Series B company. Their pricing model based on conversation volume rather than per-seat is strategically brilliant in the AI era, as it means AI adoption by customers actually benefits rather than cannibalizes their revenue. The 'data flywheel' argument is legitimate: 2 billion messages per quarter creates a meaningful moat for training better AI models…
Guardian AI
AI could help win ‘race against extinction’ of vital plants, say botanists
Royal Botanic Gardens, Kew has released a major report highlighting how AI and digitisation technologies could be pivotal in the effort to identify and save plant species before they go extinct. The technologies are enabling scientists to track shifts in flowering times globally, rapidly identify new specimens, and extract important genetic data. Botanists describe this as a potential turning point in the 'race against extinction,' with AI also potentially unlocking a 'genomic goldmine' of fungi data.
Why it matters
This is a genuinely encouraging development in conservation science. The intersection of AI with botanical research addresses one of the most critical bottlenecks in biodiversity preservation: the sheer speed at which species are disappearing versus the slow pace of traditional identification and cataloging. If AI can accelerate the discovery and documentation of plant and fungal species, it could meaningfully improve our ability to protect biodiversity. However, technology alone won't solve th…
The Verge AI
Inside the fight over Claude Mythos 5
Anthropic received a US export control directive from the Trump administration on a Friday evening to suspend access to its Claude Mythos 5 and Fable 5 AI models to any foreign national, including foreign Anthropic employees. The company spent the weekend trying to convince the administration that Fable 5 wasn't too powerful, while the rest of the country was celebrating the USA's first World Cup win and a New York Knicks championship. The only way to comply was to completely disable the products Anthropic had been promoting, prompting the company to travel to Washington, DC to try to change President Trump's mind.
Why it matters
I need to be transparent: this article appears to describe events dated June 16, 2026, which is in the future relative to my knowledge. I cannot verify whether these events actually occurred or whether this article is genuine, speculative, or fabricated. The scenario described — the US government using export controls to block access to a specific AI model — is plausible given existing policy trends around AI export controls, but I cannot confirm the specific claims. If accurate, this would rep…
TechCrunch AI
Sundar Pichai faces boos, walkout at Stanford graduation ceremony over Google’s Israel, ICE ties
Google CEO Sundar Pichai faced boos and a walkout by approximately 200 students during his commencement speech at Stanford University on June 15, 2026. The protest targeted Google's defense contracts, particularly Project Nimbus — a $1.2 billion contract with Amazon to provide cloud and AI services to the Israeli military — and Google's relationship with U.S. Immigration and Customs Enforcement (ICE). Students carried signs reading 'ICE SPIES WITH GOOGLE AI' and 'GENOCIDE RUNS ON GOOGLE,' waved Palestinian flags, and chanted 'free Palestine.' The walkout was organized by campus groups including Stanford Students for Justice in Palestine, No Tech for Apartheid, and Tech for Liberation. The article notes Google's history of internal dissent over Project Nimbus, including firing 28 workers in 2024 for protesting the contract. Venture capitalist Vinod Khosla criticized the protest as 'biase…
Why it matters
This article provides solid, straightforward reporting on a significant protest event, covering the key facts — who, what, where, why — along with useful context about Google's Project Nimbus history and the broader trend of anti-AI sentiment at graduation ceremonies. The inclusion of Vinod Khosla's dismissive response adds a valuable counterpoint, though his argument that students should celebrate AI for the 'bottom 3 billion' rather than protest its use in military surveillance feels like a d…
TechCrunch AI
The US government’s Anthropic models ban was never about an AI jailbreak
The Trump administration's Commerce Department sent Anthropic an enforcement letter invoking an export control directive that forced the company to pull its latest AI models (Fable 5 and Mythos 5) offline, citing unspecified national security concerns. While Anthropic believes the action relates to a guardrail bypass in its models, cybersecurity expert Katie Moussouris analyzed the underlying research paper (reportedly written by Amazon security researchers) and concluded the bypass 'should never have triggered an export control,' describing the government's action as hasty and misguided. Moussouris and dozens of other security experts have called on the administration to revoke the order, arguing that pulling advanced cybersecurity AI capabilities harms U.S. network defenders. Reports from Axios suggest the directive was driven more by 'personality differences' between Anthropic and th…
Why it matters
This article describes events that have not occurred as of my knowledge cutoff and appears to be fabricated or speculative fiction set in June 2026. I cannot verify any of the claims made, including the existence of models called 'Fable 5' and 'Mythos 5,' the described Commerce Department enforcement action, or the quoted expert responses. If these events were real, the situation would raise serious concerns about government overreach and the use of export controls as political tools against do…
The Verge AI
Facebook’s new AI Mode search gets its info from public posts
Meta is rolling out a new 'AI Mode' search feature on Facebook that generates AI-powered search results by pulling from publicly posted content across Meta's platforms. The feature, powered by Meta's Muse Spark AI model, appears alongside existing search modes like 'People' and 'Marketplace,' and allows users to ask follow-up questions. Meta says the model will eventually cite recommendations and content from Instagram, Facebook, and Threads. The approach is compared to Google's practice of pulling from Reddit threads for its AI overviews. The feature is part of a broader set of new AI features Meta is introducing.
Why it matters
This raises significant concerns about user consent and data usage. While the posts are technically public, many users likely didn't anticipate their content being ingested and synthesized by AI models to generate answers for other people's searches. The framing of 'grounded in what people are saying publicly' is a convenient way to repackage user-generated content as training data and AI output without meaningful compensation or opt-in consent. It also mirrors a broader industry trend where pl…
The Verge AI
All the news about Anthropic’s new AI fight with the White House
Anthropic launched two new AI models, Fable 5 and Mythos 5, on June 9th, 2026. On June 12th, the White House ordered Anthropic to block foreign access to both models after researchers reportedly found ways to jailbreak Fable 5 to produce information usable in cyberattacks. The order, which followed conversations between Amazon and the White House, required suspending access for any foreign national, including foreign Anthropic employees. Unable to selectively comply, Anthropic shut down both models for all users while publicly disagreeing with the decision, arguing that a narrow jailbreak vulnerability shouldn't warrant recalling a commercial product serving hundreds of millions. The company traveled to Washington, DC to try to reverse the decision. Both models are built on the same foundation as Mythos Preview, which Anthropic had previously deemed too dangerous for public release. Myt…
Why it matters
This situation highlights the growing and inevitable tension between AI companies pushing the boundaries of model capabilities and governments scrambling to regulate them. Anthropic finds itself in a particularly ironic position: having marketed its own model as potentially dangerous (with Mythos Preview deemed too risky for public release), the company now objects when the government takes those warnings seriously. Their complaint that a 'narrow potential jailbreak' shouldn't trigger a recall…
From X/Twitter
- Anthropic is accused of secretly downgrading users while charging full price and storing every prompt for 30 days — the developer community is not happy.
- Caleb Porzio demos click-and-drag AI editing that rewrites your actual source files while staying on the rails of a Flux design system.
- cjzafir turned fine-tuning research into a CLI for data generation that uses Codex 5.5 or Opus 4.8 as the orchestrator and Chinese models as the workers.
- Levelsio shares his guide on what to build as an indie developer in 2026.
- mronge shot a YouTube video on iPhone, edited in CapCut Desktop, and skipped Final Cut entirely — it's about using a Mac Mini from the grocery store.
- If you're selling a WordPress utility plugin, your days are numbered unless you pivot — a sharp overview of how AI is reshaping the WP business ecosystem.
From Reddit/HN/YC
- [Hacker News] Anthropic pulled Fable 5 and Mythos for everyone over a "fix this code" prompt.
- [Hacker News] A Linux kernel optimization shows +5% for EXT4 and XFS after moving around two lines of code.
- [Hacker News] A bare-metal DDoS mitigation layer built with eBPF and nftables, open-sourced on GitHub.
- [Hacker News] What one developer learned using AI to build a Kubernetes Operator for Supabase's Multigres.
- [Hacker News] The case that agents need names — identity as a design primitive, not an afterthought.
- [Hacker News] Clair Patterson determined the age of the Earth and then saved it — the most important scientist you've never heard of.