A high schooler's open-source project called Colibri runs a 744-billion-parameter GLM model using only 25GB of RAM by streaming weights from disk, achieving frontier-level capability on consumer hardware.
Snapshot · This Week in Tech (Audio)
A high schooler's open-source project called Colibri runs a 744-billion-parameter GLM model using only 25GB of RAM by streaming weights from disk, achieving frontier-level capability on consumer hardware.
Where this was said
At 2:41:30 · chapter starts 2:33:00
In one of the week's most legally complex stories, Waymo's remote operator spotted two 15-year-olds drinking and shooting toy guns from inside a robotaxi in San Mateo. The company disabled the vehicle and called police; the San Mateo Sheriff's department posted on Facebook, 'Parents, do you know where your teens are? Waymo does.' Police confirmed the teens could have exited before officers arrived but did not. [1] — Leo Laporte "In San Mateo, Waymo's operator detected two 15-year-olds drinking and shooting toy guns from the robotaxi, disabled the vehicle, and called…" 2:33:10 Patrick Beja notes the action feels categorically different when it's an autonomous vehicle calling the police rather than a human driver making a judgment call. Wesley Faulkner raises an unexplored legal question: does commercial video footage automatically give authorities access without a subpoena, and did the teens have a reasonable expectation of privacy in a for-hire vehicle? A separate story covers Waymo vehicles running out of power during July 4 San Francisco gridlock and being towed — and a video of a Waymo driving through fireworks. Patrick marvels that autonomous vehicles are genuinely working at scale while the public has largely stopped caring, citing the classic Gartner hype cycle.
In San Mateo, Waymo's operator detected two 15-year-olds drinking and shooting toy guns from the robotaxi, disabled the vehicle, and called police. Wesley Faulkner raises the unanswered legal question: does Waymo's commercial video feed give authorities automatic access without a subpoena, and did the teens have an expectation of privacy?
A young developer's open-source Colibri project runs a 744-billion-parameter GLM model using just 25GB of RAM by streaming weights from disk. It's only 4 tokens per second and CPU-only, but it proves that frontier-level local AI on consumer hardware is a real direction — especially relevant if China restricts access to its open-weight models.
Ad-based monetization works well for game apps where users spend extended time in-session, as seen with Grid and Wordle.
Tool-focused apps like PuffCount are poor candidates for ad monetization because users don't stay in-session long enough.
A hard paywall is a screen that blocks all app features unless the user pays or starts a free trial — it cannot be dismissed.
Mobile apps are primarily monetized through either ads (best for games) or in-app purchases/subscriptions (best for tools).
According to the episode, YouTube outperforms every other social platform for building trust and driving SaaS conversions.
Vasco stated that the majority of his app's user base came directly from his YouTube channel.
SEO Bot features a 'Boost My Domain Rating' button that routes users directly to Listing Bot, an example of in-product cross-selling.
The founder's entire product portfolio is AI-related, making it easier to package products attractively for directories.
The founder attached their SaaS demo to the trending debate about whether AI coding is actually good enough to build a full SaaS product.
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.