Welcome to OHMO AI. This week Washington reached in and switched off a model overnight, the labs lined up for the public markets, and a support bot quietly became a break-in tool. The systems are powerful enough now that governments, investors, and criminals all want a piece.
In today’s newsletter:
The US forces Anthropic to cut off access to its top models
OpenAI files confidential paperwork that points toward an IPO
Perplexity holds firm on a 2028 listing as rivals race ahead
A Meta AI support tool gets abused to hijack 20,000 Instagram accounts
LATEST IN AI
ANTHROPIC
US forces Anthropic to disable Fable 5 and Mythos 5
Image source: OHMOAI
OHMO AI: Anthropic said it had to disable all customer access to its Fable 5 and Mythos 5 models to comply with a Trump administration directive barring foreign nationals from using the systems, according to Bloomberg. The company said it received the order on Friday evening and disagrees with it, calling the situation a misunderstanding it is working to reverse.
The details:
The directive put both models under export controls covering any location outside the US and all foreign persons inside it, citing national security authorities.
The government's stated concern was a method of bypassing, or jailbreaking, Fable 5; Anthropic argues a narrow potential jailbreak should not justify recalling a model used by hundreds of millions.
To comply, Anthropic said it had to abruptly cut off access for everyone, including US customers and its own foreign-national employees, while other models stay live.
Why it matters: This is the most aggressive move yet to treat a frontier model like a controlled weapon rather than software. A government flipping off a commercial AI overnight is a new kind of risk for anyone building on these tools. Where you live may soon decide which AI you're allowed to use.
PRESENTED BY VIKTOR
From idea to shipped tool in 11 minutes.
Type the problem in Slack. Viktor writes the code, deploys to your subdomain, posts the URL, and starts using it on your next request. No specs, no Jira, no kickoff. Founders are running entire companies this way.
OPENAI
OpenAI files confidential paperwork that points toward going public
Image source: OHMOAI
OHMO AI: OpenAI confirmed it has submitted a confidential draft S-1 to the SEC, the first formal step a company takes toward an eventual IPO. In its own announcement the company said it expected the filing to leak, so it got ahead of it.
The details:
OpenAI says it has not decided on timing, and that going public could be a while off because some plans are easier to pursue as a private company.
The filing gives the company the option to go public sooner if that turns out to be the better path.
The announcement was made under a securities rule that lets a company confirm a submission without it counting as an offer to sell shares.
Why it matters: A confidential S-1 is the clearest signal yet that the biggest name in AI is thinking seriously about public markets. Filing keeps the door open without committing. When a company this central starts prepping for Wall Street, the whole industry's math changes.
PERPLEXITY
Perplexity sticks to a 2028 IPO as rivals rush the markets
Image source: OHMOAI
OHMO AI: Perplexity is still planning to go public in 2028 no matter how the Anthropic and OpenAI listings land, CEO Aravind Srinivas told CNBC, in comments reported by Reuters. The remarks came as OpenAI confirmed its confidential filing and Anthropic's own filing landed the week before.
The details:
Srinivas said the company has stuck to 2028 as its earliest IPO date, which Chief Business Officer Dmitry Shevelenko said has helped it build a healthy, high-growth business.
He said this week's SpaceX IPO would be a leading indicator for how the Anthropic and OpenAI listings might be received.
Srinivas added that he expects the AI IPOs to go well because the companies are doing well, while admitting there would be ripple effects if they don't.
Why it matters: While two rivals sprint toward the public markets, Perplexity is betting that patience reads as strength. The next few listings will set the price of confidence for the whole AI sector. Everyone is watching who blinks first.
PRESENTED BY WISPR FLOW
Your best prompts are the ones you'd never bother typing.
The detailed ones. The ones with examples and edge cases. Wispr Flow lets you speak them instead — clean, structured, ready to paste into any AI tool. Free on Mac, Windows, and iPhone.
META
Meta AI support tool abused to hijack 20,000 Instagram accounts
Image source: OHMOAI
OHMO AI: Meta confirmed that 20,225 Instagram accounts were potentially compromised after attackers abused its AI-powered High Touch Support system to send password reset links to email addresses they controlled. The flaw, disclosed on June 8, let strangers take over accounts that had no two-factor authentication.
The details:
The recovery tool failed to check that the email submitted with a reset request actually matched the account on file, so it sent the link to an attacker's inbox instead of rejecting it.
The flaw was exploitable from April 17 until Meta pulled the tool in early June, and the company says it discovered the issue on May 31.
Accounts with two-factor authentication enabled could not be taken over, even after a password reset.
Why it matters: This is the textbook warning for every company racing to wire AI into customer service. An assistant built to help locked-out users became the fastest way to lock them out. If you use Instagram, turn on two-factor authentication today.
AI TOOLS
Useful AI Tools
Base44: No-code AI app builder that turns your words into fully functional apps in minutes.*
Chatbase: Build and deploy custom AI support agents for your business, no coding needed.*
ElevenLabs: Generate realistic, studio-quality voiceovers from text in any language or voice.*
Magica: All-in-one AI platform with 5000+ tools for text, images, videos, audio, and lots more.*
Kick: Automates bookkeeping and tax preparation for business owners using AI.*
WisprFlow: Voice-to-text AI that turns your speech into polished writing, so you can just talk instead of type.*
*Sponsored Listing
COMMUNITY
Featured Reader
Every newsletter, we feature how a reader is using AI to work smarter, save time, or make life easier.
Today's reader is Marisol R. from Tucson, Arizona:
“I run a two-chair hair salon and I'm dyslexic, so writing confirmation texts and replying to booking messages always took me forever and I'd second-guess every word. Now I keep a chat open and type how I'd say it out loud, then ask it to clean up the spelling and tighten it. My clients get clear messages in seconds and nobody knows I spent twenty years dreading the keyboard. It gave me back the part of running my own place that used to make me want to hide.”
Want to be in the next newsletter? Tell us how do you use AI here.
AI Meme
Go Deeper
AI Tools: A curated list of the AI tools worth knowing.
Instagram: The AI page they pray you’ll never find.
