Is AI Profitable Yet? — The Question That Drew 60 Points on Hacker News
A simple question — "Is AI Profitable Yet?" — has sparked debate on Hacker News, racking up 60 points and 23 comments.
Launched by isaiprofitable.com, the project attempts to track the profitability of AI companies in real time, offering an overview of the financial status of major players such as OpenAI, Anthropic, and Google DeepMind. As of 2026, the cost of developing frontier models has swelled to billions of dollars, and the gap between that spending and revenue now stands out as an industry-wide challenge.
Weighing in on the debate from another angle is ReadUncut's report, "94% companies will keep spending on AI even when it fails." Even as many AI projects fail to deliver the expected results, an overwhelming majority of companies say they intend to keep investing. This has given rise to a critical reading: that AI investment is driven less by rational judgment than by FOMO — the fear of being left behind.
At this point, the landscape appears to be a mix of areas with clear ROI in specific use cases (code generation, customer support automation, and the like) and areas where the cost-benefit picture is still murky.
Sci-Hub Builds an AI Chatbot — A New Development in Academic Access
C&EN (Chemical & Engineering News) has reported that Sci-Hub, the site known as the pirate repository of academic papers, has developed an AI chatbot.
The chatbot appears designed to answer users' questions based on the contents of papers, drawing on Sci-Hub's vast accumulated database. Academia, however, has responded with sharp criticism, asking whether AI should now be built on top of copyright infringement. At the same time, some point out that it could prove a useful tool for researchers in developing countries, where opportunities to access papers are limited.
Sci-Hub's move sits at the intersection of the open-access debate in scholarly publishing and the copyright questions surrounding AI — a point worth watching. A full-scale evaluation of its quality is expected to get underway soon.
Google AI IDE "Antigravity" Forced Update Sends Shockwaves Through the Developer Community
A forced update to version 2.0 of Google's AI IDE "Antigravity" is causing an uproar in the developer community.
According to a detailed report on Zenn, on the morning of May 20, 2026, users who launched the IDE found that the file explorer, terminal, source control, and editor had all vanished, leaving nothing but a single chat input box. It was a fully forced update — no advance notice, no option to roll back to the previous version.
The story drew a massive response on Hacker News, scoring over 600 points and raising a fundamental question: are developers losing control over their own tools? As Google's AI strategy tilts ever more toward "chat-first," criticism is mounting that the company is disregarding existing developer workflows.
Giving AI Agents a Wallet — The Agent Wallet Vision
Attracting attention on Zenn is the concept of the "Agent Wallet": giving AI agents an independent wallet (payment capability) of their own.
"kawasekit," an OSS proposed by k0yote, is a project to build an SDK over six months that would let AI agents make payments autonomously using stablecoins. Leveraging Web3 backend expertise, it aims to provide the foundation for AI agents to automate payments for API usage and micro-payments between services.
The claim that AI agents need "economic self-determination" in order to act autonomously deserves attention as a foundational technology for the agent economy. That said, the security and governance challenges remain significant.
Search Engines vs. AI — Ripples from Nikkei's Report That Half of Twenty-Somethings Don't Use Search Engines
Following a Nikkei article reporting that roughly half of people in their twenties no longer use search engines when looking things up, the "search engines vs. AI" debate has grown lively in Japan's tech community as well.
An analysis on Zenn points to what this debate overlooks. Google's search results themselves have made "reaching accurate information" difficult, owing to SEO-driven optimization and snippet displays — the argument being that AI and search engines each carry their own structural limitations.
What matters is not "which tool to use" but cultivating the ability to verify information itself. More than a debate over the merits of particular tools, this is a discussion that has once again thrown the need for information literacy education into sharp relief.
---
Is AI Profitable Yet? – isaiprofitable.com (Hacker News)
% companies will keep spending on AI even when it fails – ReadUncut
Sci-Hub has created a new AI chatbot. Is it any good? – C&EN
GoogleのAI IDEが勝手にチャットボットに変わった——Antigravity 2.0強制更新騒動の全容 – Zenn / lumichy
AIエージェントに財布を持たせる: なぜ今、Agent Walletが必要か – Zenn / k0yote
検索エンジンvsAI論争が見落としていること – Zenn / 増田