Techin' Geek

Anthropic published a blog post Friday seeking to answer some basic questions about how it will watermark the text generated by its chatbot Claude. Such as: How will the watermarking actually work? Can it be hidden with editing? And how does this affect code?

Claude users have been debating the move since the company revealed earlier this week that it would be doing this watermarking to comply with the EU AI Act’s Transparency Code, which requires AI companies to use systems that make it possible to identify AI-generated content.

On Reddit, for example, one poster characterized this as a conspiracy against innocent Claude users, while another claimed, “The only reason you wouldn’t want this is to lie to people.” And Business Insider reports that “dozens” of users on X have claimed to cancel their Claude subscriptions as a result.

Anthropic’s new post starts with a general overview of the watermarking concept, explaining that when making “low-stakes choices” — like choosing between the words “overcast” and “grey” to describe the weather — Claude can create a pattern in its responses that is “undetectable to the reader, but is detectable to anyone who has a key that encodes it.”

“Watermarking does not impact the quality of Claude’s output,” the company said. “To a reader, a watermarked response is indistinguishable from an unwatermarked one.”

More specifically, Anthropic said it will be using the SynthID-Text approach that the Google DeepMind team outlined in 2024, and that it plans to release a watermark detection API. It also noted that watermarking is distinct from the AI detection approaches offered by companies like Pangram that look for “tells” in the writing (like the construction “his isn’t [X], it’s [Y]”) to reveal AI usage: “Picking up on these patterns is fundamentally different from checking for a watermark.”

Could someone just rewrite the text to hide the watermark? Anthropic said it’s possible, but “light editing probably won’t remove the watermark completely,” while “a complete rewrite where every word is replaced will.”

“In the latter case, of course, it’s arguable whether the text can any longer be described as AI-generated,” the company said.

As for whether the watermark will be detectable in text that was only proofread or edited by Claude, Anthropic said that will depend on “the length of the text and how heavily Claude has edited it.” If it’s only been lightly edited, “nearly all the words” will have been written by the human author and “there’s very little (if anything) for the watermark to attach to.”

Code, meanwhile, should have less of a watermark than other text, because the model will need to create working code and won’t have the freedom to choose between a variety of equally valid options. 

“Having said that, in areas where there is an arbitrary choice between particular words or terms within the code, the watermark can be used, such as comments within code,” Anthropic said. “But by definition, it will have a negligible effect on the actual code produced.”

Anthropic also said that Claude won’t be the only AI chatbot to generate watermarked text, as “other major model developers have signed the same Code of Practice and will be implementing their own watermarks.”

Just like any other online service, hackers can target and break into your accounts on popular AI platforms such as ChatGPT, Claude, and Perplexity. 

TechCrunch has created a comprehensive guide to help you protect yourself if you suspect someone has broken into your account on one of the internet’s most popular platforms, social networks, or messaging apps. Now, we’re here to show you how to check whether your accounts on AI platforms have been hacked.

As usual, we recommend using unique passwords stored in a password manager, and turning on multi-factor authentication (MFA), so that even if someone steals your password, they won’t be able to log in without that second piece of information. 

ChatGPT and Perplexity offer MFA. Claude doesn’t, because instead of asking for a password, Anthropic’s AI chatbot sends a login link to your email address.

All three of these AI platforms offer similar ways to check if there’s a suspicious device logged into your account. Here’s exactly how each platform works.  

ChatGPT

To find out if someone has broken into your ChatGPT account, open it on your computer’s browser, click on your username in the bottom left corner, go to “Settings,” then “Security and Login,” and finally click on “Active Sessions.” 

You will see where you are logged into your ChatGPT account. If you see any device you don’t recognize, you can log out of that single device. You can also click on “Log out all.”

At this point, if you want to change your password, you need to log out of your account. 

Then, on ChatGPT’s website, click “Log in” located in the bottom-left corner, enter your email address, click on “Forgot password,” and then “Continue.”  

ChatGPT will then send you an email containing a six-digit code. Enter the code on the ChatGPT login page, click “Continue,” and then enter a new password. 

You can also click “reset your password” in the email you received to see the official instructions on how to do that. 

Claude

For Claude, open it in your computer’s browser, click on your username in the bottom-left corner, then “Settings,” and click on “Account.” That’s where you will see your “Active sessions.” 

If you don’t recognize one of them, hover over it, click on the three vertical dots that appear on the right, and click “Log out” or “Terminate.”

If you want, you can click on “Log out of all devices.” 

At that point, you’ll be able to log back into your account using your email address. You will receive an email with a link to log in. Claude does not allow you to use passwords at all, so there’s no password to change.

Perplexity

In the case of Perplexity, the AI-powered search engine does not show you where you are logged in. 

So if you’re worried someone may have broken into your account, go to Perplexity in your browser and click your username in the bottom-left corner, then “All settings.” Finally, click on “Sign out of all sessions,” and then “Confirm.”

At that point, you can log back in by entering your email address. You will then receive an email with a unique six-digit code. Enter the code on the website to log in, or click on the “Sign in” button in the email to log in directly. 

Ceva Logistics, a major player in the global shipping and logistics sector, has experienced a cyber breach. Various companies depending on Ceva for their shipping needs report that their sensitive data was compromised in the attack.

The cyber breach has impacted at least eight distribution centers throughout Europe involved in the shipment of products across the region, as stated by the company to TechCrunch.

According to industry publication FreightWaves, the attack commenced on July 29, leading to delays in the shipment of numerous items housed in the affected facilities.

Ceva, headquartered in France, serves as a critical shipping and logistics provider for businesses globally, facilitating the distribution of goods from production sites to consumer residences. The firm reported $18.3 billion in revenue for 2025 and operates more than a thousand warehouses worldwide.

Recently, logistics and shipping firms have increasingly become targets for cybercriminals, given their access to and control over trucks and containers laden with merchandise, often diverting it to criminal enterprises.

The breach at Ceva also led to a significant data theft, impacting a vast array of personal details belonging to retail clients that Ceva uses for home delivery services. Numerous companies disclosed that hackers accessed their customers’ names, residential addresses, phone numbers, and email addresses stored within Ceva’s systems.

The Dutch e-commerce powerhouse Bol announced on its site that intruders accessed its warehousing partner, Ceva, and cautioned that their clients’ personal data may have been compromised. Bol anticipates delays and potential cancellations of certain customer orders due to the incident.

In a similar vein, De Bijenkorf, a Dutch luxury retailer, confirmed shipping delays as a result of the theft of its clients’ data, according to local news outlets. Additionally, football club Ajax, banking entity ING, and eyeglass manufacturer Ace & Tate indicated that their customers’ shipping details were also impacted. 

Valve, the video game industry titan, informed its customers on August 7 that data had been extracted from Ceva’s systems, notifying those who recently purchased its Steam hardware that their personal information was compromised in the breach. In a message to customers shared on Reddit, Valve indicated that Ceva retains their shipping and delivery data for 90 days post-order.

Valve representative Doug Lombardi did not reply to a request for comments regarding the incident. 

Ceva acknowledged the cyberattack in a statement to TechCrunch.

“On Aug. 1, CEVA Logistics informed affected clients that a cyber intrusion was disrupting part of its European contract logistics functions. Once the breach was detected, CEVA’s cybersecurity teams promptly initiated their security protocols and commenced a comprehensive investigation that is still ongoing,” as per the statement shared with TechCrunch. “The operational ramifications are confined to eight warehouses. No other CEVA systems globally were impacted, and all remaining operations proceed without issue.”

Ceva spokesperson Ryan Fisher declined to answer TechCrunch’s inquiries about the breach, including whether the company is aware of the extent of personal data taken or if there has been any communication from the attackers, such as a ransom request.

Ceva reported that some of its affected applications and services have been restored, and that it is collaborating with law enforcement. At the time of publication on Monday, Ceva’s website was not loading properly.

Investigative efforts by authorities in the Netherlands are underway regarding the incident. 

Mark Schenkel, a representative for the Dutch data protection authority, did not respond to TechCrunch’s request for comment on Monday concerning the situation.

Recent security findings indicate that marketing technology leader Klaviyo was, until recently, unintentionally disclosing the sign-up details of its new users, including their passwords, to external advertisers.

Sam Jadali, a security investigator and co-founder of the cybersecurity company Melurna, informed TechCrunch that the web form on Klaviyo’s registration page was incorrectly set up from at least February 2024 to November 2025, likely for an even longer period.

The startup’s investigations revealed that anyone registering with Klaviyo via the faulty form might have had their sign-up details shared with various third-party tech companies and advertisers whose tracking mechanisms are also integrated into the company’s website.

This sign-up information contained the user’s email, password, as well as the organization’s name, website, and phone number. This data was disclosed to advertising and tech firms such as Facebook and Google; marketing leader HubSpot; Microsoft and its subsidiary LinkedIn; social media platform X, among others.

The startup disclosed its results to TechCrunch prior to its presentation at the Def Con security conference in Las Vegas.

Klaviyo acknowledged to TechCrunch that it rectified the website flaw, yet uncertainties remain regarding the incident, including the total number of individuals impacted by the data exposure throughout the years. The Boston-based marketing leader permits its 205,000 paid clients to execute advertising campaigns through email, SMS, and various other means. Klaviyo’s site claims it oversees over seven billion customer profiles.

The flaw highlights the data vulnerabilities that third-party trackers can inflict on website users when defensive measures, like ad-blockers, are not employed. Klaviyo joins the list of companies in recent times that have unintentionally exposed data to external entities.

Website trackers, referred to as “pixels,” enable website and app proprietors to gather information concerning their visitors and users, commonly for understanding how their applications are utilized and for detecting issues. These trackers can be misconfigured to also disclose personal information entered on any webpage where they are present. 

In recent years, security breaches arising from improperly configured pixel trackers have led to businesses submitting data breach notifications and regulators initiating enforcement measures.

When contacted by TechCrunch, Klaviyo representative Danielle Zanatta confirmed the flaw was linked to an “application configuration issue.” Zanatta mentioned that the count of known affected individuals was fewer than 200, “based on our readily available active logs.” Klaviyo refrained from commenting on how far back it retains logs or how long the flaw was present on its website.

Klaviyo stated it informed the known affected individuals, but did not provide a copy of the communication that the company supposedly sent to impacted customers when requested by TechCrunch. 

It remains uncertain why the company did not disclose the incident publicly. 

Chips executing AI tasks tend to overheat: This is one factor contributing to the high electricity usage in data centers and the need for cooling solutions. Consequently, innovators are leveraging AI to address the issue it has caused.

Discovered Materials is the newest entrant, aiming to employ swarms of AI agents to uncover new materials that enhance the creation of more efficient integrated circuits. The startup recently announced it has secured a $9 million seed funding round from Lightspeed India Partners, having recently graduated from Y Combinator, with investments from Peak XV Partners and angel investors such as Paul Graham, Gokul Rajaram, and Thariq Shihipar.

Founders Advaith Sridhar and Akash Ramdas joined forces to establish the company, leveraging Ramdas’ expertise from earning a PhD in materials science at Stanford and Sridhar’s experience with agents at Persona AI and Luma Labs.

The duo has developed a software pipeline that utilizes Anthropic models within a custom setup to generate leads for materials, then applies foundational physics models they have trained to conduct simulations that confirm the potential of candidate materials.

“[Ramdas] was making perhaps 20 hypotheses a day during his PhD,” Sridhar shared with TechCrunch. “Now, we can generate thousands of hypotheses daily by having these agents operate round the clock in the cloud, exploring the research directions he provides.”

Discovered Materials unveiled examples of hundreds of new materials today, along with their “Material Discovery Bench,” designed to monitor how cutting-edge models tackle this challenge.

Similar initiatives have emerged from companies like MatNex, SandboxAQ, and CuspAI, but Discovered Materials is focusing specifically on the thermal challenges faced by semiconductor materials as a pathway to success. The startup claims it has already identified several materials with properties comparable to those used by leading chipmakers but is unable to disclose further specifics.

One of the hurdles is the engineering trade-off: If a new material could lower heat generation or enhance dissipation, it may still be impractical to manufacture a chip using it, or its electrical characteristics might be diminished.

“It’s somewhat like playing whack-a-mole with atomic structures,” remarked Hemant Mohapatra, the Lightspeed partner who spearheaded this funding round, in conversation with TechCrunch. “A material is only practical in the real world if all properties align simultaneously, which adds intrigue to this search problem.”

Mohapatra anticipates that the market for predicting new substances will become commoditized as models advance. The distinction with Discovered Materials lies in Ramdas’ extensive experience in the domain and the capability to operate a lab capable of swiftly experimenting and validating potential candidates — a feat he asserts both founders have already accomplished with multiple new materials.

Once valuable candidates are identified, Sridhar mentions the company will seek to patent the application of these materials in GPUs, or the method of creating chips from the substances, licensing them to chip manufacturers. He aspires to have patent-worthy new materials within the year.

However, amidst all the enthusiasm, there has yet to be any drugs or materials discovered by AI that have made a significant commercial impact. The closest instance might be Insilico Medicine’s Renterosib, the inaugural drug found with generative AI that has progressed to a Phase II clinical trial. In terms of materials, promising contenders like MatNex’s rare-earth-free permanent magnets and new semiconductor materials developed by Panasonic and Citrine Informatics have been identified, but they have not yet been deployed commercially at scale.

These approaches may be maturing as AI technology progresses, yet Mohapatra contends that finding more candidates is not the bottleneck in AI materials science; rather, “the challenge lies in accurately filtering and synthesizing them.”

While Sridhar believes Discovered Materials’ unique data and insights will enable the startup to compete with well-funded advanced labs, he acknowledged that much of the work will involve actually going into wet labs to create materials. This is a phase of the process that cannot be expedited.