Adobe Launches AI Assistant in Photoshop Public Beta

Lights on, Tech Insiders.

From Photoshop's voice-activated magic to billion-dollar data fortresses and malware-laced copy-paste installs, innovation is shining bright and revealing some shadows.

Follow the glow before it blinds you.

Here's what you need to know today:

Photoshop's AI Assistant Hits Public Beta

Who needs 50 layers when you've got one chatbot?

Yesterday, Adobe's new AI Assistant for Photoshop entered public beta on the web and mobile apps (sorry, desktop loyalists, you'll have to wait).

Type or speak requests like "remove that photobomber" or "make the sky neon pink," and the assistant will do the heavy lifting or walk you through the steps.

A companion AI Markup tool (currently web-only) lets you scribble on an image to pinpoint exactly where changes should appear, then uses generative magic to fill the gaps.

Mobile users get voice commands for edits on the go, while web users can watch each step unfold in the History panel (handy if the bot's creative vision needs a human tweak). Through April 9, paying subscribers enjoy unlimited AI generations; free users start with 20.

Firefly isn't left out, with its revamped Image Editor bundling a full suite of generative tools (Fill, Remove, Expand, and Upscale) alongside one-click Background Removal into a single workspace. Creators can now tap more than 25 AI models, including Google's Nano Banana 2 and OpenAI's Image Generation, without breaking their flow.

Why it matters: Photoshop's learning curve just flattened. Whether you're a pro polishing client work or a hobbyist fixing vacation pics, natural-language edits free you from menu mazes and let you iterate faster. Plus, unlimited generations (for now) mean you can experiment wildly without sweating credits. Creative block, meet your new nemesis.

Share this newsletter!

Pulse Check

Will Adobe's AI assistant change how you edit photos?

Results from Yesterday's Pulse Check

Would you let Copilot Cowork rearrange your week?

Insider Intel

Nscale Secures $2B for Nvidia AI Data Hubs

Turns out the hottest land grab isn't beachfront; it's GPU acreage.

Earlier this week, UK-based Nscale announced a $2 billion Series C co-led by Norway's Aker ASA and 8090 Industries, with Nvidia, Dell, and others, valuing the startup at $14.6 billion. It's the largest Series C in European history.

Nvidia's check signals the chip king wants to own not just the silicon but also the racks those chips call home.

Now fully absorbing its Aker joint venture, the cash lets Nscale accelerate a global web of AI-optimized data centers—liquid-cooled, renewably powered, and packed with hundreds of thousands of Nvidia accelerators.

Flagship projects include Stargate Norway (100,000 GPUs for OpenAI) and a $14 billion Microsoft pact to host 200,000 Nvidia GB300 accelerators across Texas, Portugal, and the UK.

For enterprises racing to train and deploy models, that means another well-capitalized supplier of scarce compute, with pay-as-you-go or dedicated racks spun up in weeks instead of years.

Nscale's vertical control, from power contracts to orchestration software, could pressure hyperscalers on price and speed, giving smaller firms a way around the GPU waitlist.

Board additions Sheryl Sandberg, Nick Clegg, and Susan Decker arm the 2024-born company with policy and governance firepower ahead of a rumored IPO.

Consumers won't care about the plumbing, at least until their apps suddenly get faster, cheaper AI features drawn from these megawatt machines. Of course, if the AI bubble deflates, those shiny server farms could end up subletting compute to cat-video renderers.

If data centers are the new oilfields, former coal-miner-turned-CEO Josh Payne just struck Texas tea, digitally speaking.

Flash Bytes

Secure Your Seat for Compliance 2026

Compliance effort is rising, but audit confidence isn't.

Join The State of Audit and AI Compliance in 2026 on March 19 at 1:00 p.m. ET for a research-backed discussion led by Thoropass. The session will examine where compliance programs are struggling, why audit execution has become the primary risk, and how leaders are restoring governance and visibility.

Security Alerts

Fake Claude Code Installs Spread 'InstallFix' Malware

Cloned Claude Code install pages are topping Google ads, pushing one-liner curl commands that fetch the Amatera Stealer, a data-vacuuming infostealer.

In this "InstallFix" attack, threat actors clone Anthropic's Claude Code docs, swap the URL, and use trusted platforms like Cloudflare to dodge filters and hijack developer terminals.

Protect yourself by skipping sponsored ads, using official docs, and reading URLs before piping to bash. Better yet, type commands manually. Add DNS filtering, monitor mshta/cURL chains with EDR, and restrict admin rights.

Because nothing says "move fast and break things" like gifting strangers your session tokens and SSH keys.

Whistleblower: SSA Data Walked Out on Thumb Drive

Social Security's inspector general is investigating claims that a former Department of Government Efficiency (DOGE) engineer smuggled the agency's Numident and Master Death File databases—records on more than 500 million living and deceased Americans—onto a personal thumb drive for his new private-sector gig.

If confirmed, this unprecedented insider breach exposes SSNs and birth dates to unknown misuse.

Worse? The engineer allegedly bragged to colleagues about keeping his "God-level" access post-departure and expecting a presidential pardon if caught.

Defend your perimeter: Audit admin privileges, disable removable media, and monitor for anomalous data pulls. Supply-chain partners must verify they haven't ingested stolen records.

Industry Shakeups

OpenAI Buys Promptfoo to Bulletproof AI Coworkers

OpenAI is folding security startup Promptfoo into its enterprise Frontier platform, betting that ironclad testing will persuade risk-averse companies to unleash autonomous "AI coworkers."

Promptfoo's toolset automates red-teaming for prompt injections, jailbreaks, data leaks, and tool abuse, while logging results for auditors and compliance teams.

Founded in 2024, backed by heavyweights like a16z and Insight Partners, and already used by over a quarter of the Fortune 500, Promptfoo offers both an open-source CLI and enterprise testing tools. Cofounders Ian Webster and Michael D'Angelo will keep the OSS project alive even as they bake deeper safeguards directly into Frontier's agent workflows.

For OpenAI, the acquisition plugs a critical gap as generative agents move from demos to production systems that tap real customer data and back-end APIs.

No price tag was disclosed, but Promptfoo's last valuation landed between $85.5 million and $119 million, depending on the source—pocket lint in Sam Altman's recent acquisition spree, which already swallowed up healthcare startup Torch, Software Applications (maker of the Mac AI interface Sky), and OpenClaw creator Peter Steinberger. OpenAI notes the acquisition is pending customary closing conditions.

If ChatGPT is the new intern, Promptfoo is the cranky CISO glaring over its shoulder.

Meet Our Author

Justin Meyers

Writer at TechnologyAdvice

Justin Meyers is an investigative writer and editor who draws on over a decade of meticulous hands-on research to deliver the full, trustworthy story behind consumer and enterprise tech, including cybersecurity.

Curious about where AI is really headed?

The Neuron cuts through the noise to bring you smart, hype-free takes on the latest AI trends, tools, and breakthroughs. Join 500,000+ professionals from top companies like Microsoft, Apple, Salesforce and more.

Advertise in Daily Tech Insider!

Daily Tech Insider is a TechnologyAdvice business.

© 2026 TechnologyAdvice, LLC. All rights reserved.

TechnologyAdvice, 3343 Perimeter Hill Dr., Suite 215, Nashville, TN 37211, USA.