A New Chinese Model Beat Hundreds of Human Teams in an Online Coding Competition.

Your next business idea could be one page away.

Whether you're a creator, freelancer, entrepreneur, or side hustler, AI is opening new opportunities to earn more without starting from scratch.

Discover 200+ curated AI-powered income ideas across content creation, eCommerce, marketing, automation, and digital services—all designed to help you take action.

3 things that happened while you were busy

1  Alibaba released Qwen-3.8Max, an open-weight model it calls an "always-on workmate," and it beat hundreds of human teams in an online coding competition.

The Chinese lab dropped Qwen-3.8Max, an open-weight model rivaling US frontier performance on knowledge work, coding, and long-horizon planning. Alibaba claims the model improved on an existing research paper in 125 hours, coded autonomously for more than 10 days straight, and beat hundreds of human teams in an online competition. Try it on QwenCloud or read the lab's guide to designing effective prompts.

2  A new disclosure shows ChatGPT accounts for nearly 90% of AI spending across the US House of Representatives, as major labs head to the White House today.

OpenAI's ChatGPT accounted for nearly 90% of all AI spending for members of the House, with Anthropic's Claude a distant second. The disclosure lands the same day OpenAI, Anthropic, Meta, and Google are set to meet with the White House to review a proposed framework for testing and launching new models. A lawmaker preference disclosure and a regulatory framework meeting landing on the same day is not likely a coincidence.

3  DeepSeek's cheapest model now rivals Claude Opus 4.8 in performance while charging nearly 90 times less per output token.

DeepSeek-V4-Flash charges 28 cents per million output tokens, compared to $25 for Opus 4.8, while closing much of the performance gap. US labs are responding to the sustained pricing pressure from Chinese labs: OpenAI cut prices on GPT-5.6 Luna and Terra last Thursday, and Google shipped an efficient family of Flash models earlier this month. The price war that started with a handful of Chinese labs is now visibly reshaping how every major US lab prices its cheaper tiers.

Intelligence is going the same way as oil or electricity.

The commoditization argument. AI is turning into a commodity. Any of today's top models can draft an email or build a presentation competently. Unless you are tackling a genuinely complex problem, the need to pay premium prices for frontier performance is fading fast. Zack Kass, OpenAI's former head of go-to-market, calls this "diminishing model returns," and some companies have already stopped paying the premium.

Companies are already routing around the premium. Coinbase CEO Brian Armstrong revealed in June that his company kept AI spend flat while increasing usage, largely through model routing: sending everyday tasks to cheaper models and reserving frontier models for the tasks that need them. Armstrong specifically named GLM 5.2 and Kimi 2.7 as cost-savers. Microsoft has noticed the same trend and is considering adding DeepSeek V4 to Copilot Cowork as a cheaper alternative for the tasks that do not require the top-tier model.

The load-bearing assumption. Commoditization only holds as long as cheaper models keep pace with frontier performance. If a frontier lab pulls meaningfully ahead again, companies suddenly have a reason to pay premium prices once more. Given how fast labs are releasing new models right now, that shift could happen on any given week. Commoditization is not a permanent state. It is a temporary equilibrium that holds only while the gap stays small.

What this means for how you buy AI. The strategic move right now is not picking one model and committing to it. It is building the routing infrastructure, or at minimum the habit, to move workloads between models as the price and performance landscape shifts underneath you. The companies benefiting most from this moment are not the ones locked into a single vendor. They are the ones treating model choice as a variable to optimize continuously rather than a decision made once.

What people are actually watching and sharing

😅 Alarmingly real. Flux 3 is a little too good at generating historical video footage. Its results are convincing enough that commenters are raising alarms about the tool's potential to rewrite history. Watch the clip that sparked the debate (3.5M views).

🤣 Hilariously wrong. Redditors keep asking AI to create "completely serious" images of common objects being used in the wrong context. The results are equal parts alarming and hilarious. View some of the top images. (2,000 upvotes)

🧠 Thought provoking. In one word, what is the biggest lie in human history? Someone asked ChatGPT, Claude, DeepSeek, and Gemini, and got four very different answers (2M views). The disagreement across models is itself the interesting part.

📚 Interactive education. It might take time, but AI will almost certainly reshape education. Exhibit A: one developer vibe-coded an interactive website that teaches human anatomy more effectively than any textbook (10,000 bookmarks).

🎨 Logo prompt. A viral logo prompt gives any brand logo a distinctive new twist, turning it into a holo-vinyl sticker using ChatGPT Images, great for sprucing up a website or social profile. See the original post. (5,000 bookmarks) The full prompt is in Prompt Station below.

This ChatGPT Images prompt, which picked up 5,000 bookmarks this week, takes any brand logo and transforms it into a glossy, iridescent holo-vinyl sticker with the exact rainbow-foil shine and cut-line detail of a real collectible sticker. Upload your logo as a reference image, run the prompt, and get a shareable, on-brand asset in a format that reads as premium the moment someone sees it on a laptop lid or a website footer.

COPY AND PASTE THIS PROMPT

Upload your logo as a reference image alongside the prompt rather than describing it in text, since the model needs to see the actual shape and colors to preserve them accurately underneath the holo effect. If the first result looks too generic, add one line specifying your brand's exact color: "keep the base color true to Pantone 286C blue" or similar. The die-cut border and peeling-corner detail are what sell the illusion that this is a physical object rather than flat digital art, so keep both in the prompt even if you simplify other parts of it.

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get The Free Guide