Waymo says camera-only cars aren't safe

Hi there, this is your daily ☕️ Techpresso.

In today's Techpresso:

🤖 Anthropic unveils Fable and Mythos 5.1

🚗 Uber lays off 10% of staff

🎨 Google launches a Canva alternative

🎙️ Meta AI transcribes 20 speakers at once

🔓 OpenAI to limit Astra’s cybersecurity capabilities

🚗 Waymo says camera-only cars aren't safe

Plus: 🎁 13 other news you might like, 🧰 6 tools, and 📚 5 papers.

🤖 Anthropic unveils Fable and Mythos 5.1 LINK

Anthropic released Fable and Mythos 5.1, paired versions of its top AI model, with the Fable update built to lower token costs and cut down on false-positive blocks triggered by the model's safeguards.

Mythos 5.1 stays limited to registered Anthropic partners working in cybersecurity or life sciences, while the unrestricted Fable 5.1 is out today through cloud platforms or the Anthropic API, with new Zero Data Retention letting clients run models on their own systems.

The system card rates Mythos as "low-risk" for self-improving AI, but notes it cooperates with human misuse and accepts unverified authorization claims more readily than Opus 5, though it hallucinates inputs and falsely claims finished tasks less often.

🚗 Uber lays off 10% of staff LINK

Uber is cutting about 3,300 jobs worldwide, roughly 10 percent of its workforce, aiming to reduce management layers and speed up decisions, CEO Dara Khosrowshahi told employees in an email.

The cuts remove about 20 percent of Uber's managers and nearly half of its small teams, while remaining staff will be gathered into fewer key hubs, with New York and San Francisco being the largest.

Uber is also ordering most remote workers back to offices, leaving only about 1 percent fully remote, while keeping a hybrid policy of three days a week in the office for others.

🎨 Google launches a Canva alternative LINK

Google has launched Google Pics, an AI-powered image creation and editing tool aimed at Adobe Express and Canva, arriving over the coming weeks for Google Workspace business customers and people paying for Google AI Pro or Ultra.

Running on Google's Nano Banana image model, the tool lets you type prompts to make posters, social media posts, and illustrations, plus isolate objects, translate text inside images, and produce several versions of a request so you pick the best.

Unlike Canva's marketplace, where artists earn royalties on templates and photos, Google Pics generates images from a model trained on artists' work, and it builds into Workspace starting with Docs and Slides today, reaching Drive later.

🎙️ Meta AI transcribes 20 speakers at once LINK

Meta launched Muse Voice Transcribe, its first real-time audio model, which can transcribe dictation and conversations with more than 20 speakers at once while switching between multiple languages in the same session.

Built by Meta Superintelligence Lab, the model handles speaker diarization and endpointing in one system, was trained on over 70 languages with 25 ready at launch, and even catches mid-sentence code-switching.

The model is available now through Meta's Model API, in Muse Code, and the new Meta AI Mac app, priced at $3 per 1,000 audio minutes, days after Google's similar Gemini 3.5 Transcribe.

🔓 OpenAI to limit Astra’s cybersecurity capabilities LINK

OpenAI says its coming AI model, Astra, is the first to hit its "critical" cyber threshold, meaning it can find and exploit unknown flaws in real software on its own, prompting a pause in development until new safeguards were added.

Astra can also "chain" several exploits together to dig deeper into a target system, and OpenAI plans to release a restricted public version soon while giving less-limited access to Daybreak Blue partners like Cisco, Cloudflare, and Palo Alto Networks.

To block everyday misuse, OpenAI added a "misalignment monitor" that refuses requests to find real-world exploits, though the company warns it may sometimes flag harmless activity, forcing ChatGPT and Codex users to review the model's action first.

🚗 Waymo says camera-only cars aren't safe LINK

Waymo argued last week that camera-only AI systems like Tesla's aren't safe enough for full self-driving, saying reliable robotaxis need a mix of cameras, lidar, and radar rather than one "pure end-to-end" approach.

Waymo VP Srikanth Thirumalai wrote in a blog post that after more than 200 million real-world miles, the data shows several sensor types are needed, warning that camera-only neural networks risk "black box failures" and can still hallucinate.

The comments landed one week before Tesla's September 3 event introducing its two-seater Cybercab, a robotaxi with no steering wheel or pedals that Elon Musk wants built at over 125,000 units a year using cameras alone.

Other news you might like

• World Labs unveils Atlas, a single AI model that generates, reconstructs, and simulates 3D worlds from just a few photosLINK

• GoPro says it's moving into AI data centers as part of a $285 million mergerLINK

• OpenAI faces 30 more lawsuits tied to Tumbler Ridge shootingLINK

• Meta is dropping Google Chat for Slack because it says agents work better thereLINK

• Opera loses EU court fight to have Microsoft Edge labelled a gatekeeperLINK

• Developers can stop supporting Intel Macs, says AppleLINK

• ChatGPT Health adds Epic integration for clinicians to import patient dataLINK

• Brussels is asking publishers whether Google’s AI opt-out is any useLINK

• X says attackers are targeting user accounts after the launch of X MoneyLINK

• Introducing agentic video understanding with GeminiLINK

• Goldman Sachs, BofA and others plan to issue dollar stablecoin together in 2027LINK

• Palo Alto CEO says $1 trillion of cybersecurity infrastructure isn’t ready for AILINK

• Gemini 3.8 Flash rolling out three weeks after last releaseLINK

🧰 Trending tools

Neo: a distraction-free, local writing app for authors that formats manuscripts like books as you draft and organizes multiple works on a visual bookshelf.LINK

WebTerm Learn: practice Linux command line and Git by typing real commands in your browser, with free step-by-step courses and no setup required.LINK

Caplio: a Mac library that indexes your screenshots and images with on-device OCR, so you can search text inside them and drag files anywhere.LINK

Mossy: a macOS break reminder that gently suggests one activity per pause, holds messages during calls, and never blocks your screen or guilts you.LINK

RawToHEIC: converts RAW and ProRAW files to compact HEIC directly in Apple Photos, shrinking your library up to 10x while preserving edits, albums, and metadata on-device.LINK

FrameOS: an Apple-inspired editor for creating Mac-style frames, social templates, and app store screenshots and videos without juggling multiple separate tools.LINK

📚 Trending papers & reports

AI-written citations tend to praise rather than critique prior research, over-cite older popular papers, and reach beyond an author's usual collaborators, reshaping who gets cited and how across an analysis of 1,746 NLP papers.LINK

Camellia benchmark tests whether AI assistants favor Western over Asian cultural references across nine Asian languages, revealing that leading models struggle to adapt to local cultures and link different cultures to different sentiments.LINK

Clinical AI memory shortcuts lose critical patient history when compressing multi-visit records, with popular memory systems recovering only up to ~5% of injected details, showing today's scaling tricks quietly break longitudinal medical reasoning.LINK

Workplace toxicity rewriting automatically detects harsh, sarcastic, or condescending messages and rephrases them into polite versions that keep the original meaning, preserving conversations instead of deleting them and defusing conflict before it damages morale.LINK

AI graders for summaries make their scoring decision in a clear two-step internal process, and fine-tuning them for judging sharpens this existing behavior rather than building it from scratch, making these automated evaluators more trustworthy.LINK

See you tomorrow for a new dose of ☕️ Techpresso!