Google appears to be swiveling its AI strategy in a way that could shake up an ongoing debate over how available the tech behind generative models should be to (almost) anyone who wants it.

The search giant has rolled out a set of models based on its flagship Gemini system. Gemma, as the initiative is called, notably comes with open access to the code underpinning them.

• Models come in two pre-trained sizes, both small enough to run on a laptop or desktop, along with a responsibility toolkit meant to make for “safer AI applications,” according to a company blog post.

Gemma models are far more available to developers to use than, say, the full code or training data for Google’s Gemini chatbot, but calling it “open access” doesn’t mean it is without rules: For example, Gemma comes with very specific terms of use. The move marks a departure from Google’s otherwise closed-door approach of late to AI releases and positions it in more direct competition with Meta’s Llama models, which have championed a similar, open approach.

Playing both sides? Should the code and training data used by LLMs be public? Companies like OpenAI (despite its name) and Anthropic have argued that the code behind their latest models could be dangerous in the wrong hands, as an explanation for their more closed approaches. Meta, IBM, Mistral, Databricks and others have argued that openness can spur innovation and collaboration.

Keep reading here.—PK

Old chain-drive technology can’t hold up to the increased power of electric motors in e-bikes and e-motorcycles…right? Well, not until Veer pulled up to the scene.

Veer’s carbon fiber belt drivetrains are replacing the old chain drives, improving the performance, range, and safety of these vehicles. With thousands of units shipped, 200% sales growth year over year, and over $2m already raised from investors, the electrification revolution is just getting started.

Now Veer is looking for partners to help bring Shift Drive, its innovative full transmission package for electric vehicles, to the market—reducing costs and enabling shifting under full electric power.

High-dollar investors: You can even get bonus shares depending on your investment level. Help fund the electrification revolution.

AT&T CEO John Stankey apologized Sunday for a nationwide service disruption that prevented AT&T customers from making or receiving phone calls.

In a letter to employees that was released publicly, Stankey acknowledged that the outage, while not nefarious, was frustrating and disappointing.

• “While it is not unexpected to encounter challenges as we enhance and expand our network, we have processes and redundancies in place for a reason,” he wrote. “We owe it to our customers and ourselves to do better.”

Stankey also noted that the company is crediting affected accounts with $5, an amount the company said is equal to roughly a day of service.

The network began faltering in the wee hours of Thursday morning, showing customers SOS mode, which only allows emergency phone calls, if any at all. At the height of the outage, more than 60% of customers across states including California, Texas, Florida, and New York experienced the disruption, according to data from Opensignal.

What happened? Despite some speculation that AT&T could have fallen victim to a cyberattack, the company said the error was solely an internal one.

“Based on our initial review, we believe that [Thursday]’s outage was caused by the application and execution of an incorrect process used as we were expanding our network, not a cyber attack,” according to a “network update” note on AT&T’s website. “We are continuing our assessment of [the] outage to ensure we keep delivering the service that our customers deserve.”

The Federal Communications Commission, which regulates telecom carriers, said Thursday on X that its Public Safety and Homeland Security Bureau was “actively investigating” the outage. A spokesperson declined to comment further as the agency’s probe is ongoing. A note on AT&T’s website said that the carrier had restored service to everyone affected by around 3pm ET.—KG

Most of the content on TikTok comes from the same, small group of users.

The “most active” 25% of US adult users on TikTok are responsible for 98% of public content on the platform, the Pew Research Center found in a new report published on Thursday. Pew determined the threshold for “most active” as a user who’s posted at least seven public videos.

Pew previously found the same trend on X, where the most active 20% of users were responsible for 97% of all tweets in 2021.

“Notably, there are no significant differences in the share of users who have posted on the site based on gender, political affiliation, or educational attainment,” the report found. Pew surveyed 2,745 TikTok users for the new research.

TikTok’s US growth has been explosive—last year, the company touted 150 million US users on the platform, and it’s investing heavily in its stateside e-commerce business, with mixed results.

Keep reading on Marketing Brew.—RB

Calling all innovators . Technologists at Millennium partner across the organization to build solutions that drive impact and help make the business faster + more resilient. Ready to contribute to a team that fuels innovation on a global scale? Check out Millennium’s open roles to get started.