AI News Weekly
AI developments selected from public sources
- 111
- major stories
- 34
- sources
- 44
- first-party stories
Lead story
English translation is being prepared.
newsletter
English translation is being prepared.
Ahmad Al-Dahle, former head of generative AI at Meta and now CTO of Airbnb, told Latent Space about Airbnb's AI-native transformation: currently 60% of code is written by AI, feature delivery volume has increased nearly 80% year over year, and per-engineer PR throughput has risen about 1.6x. Customer support was the first user-facing deployment scenario, with about half of tickets resolved by AI, which the company says is consistent with the nearly 45% figure in its Q2 earnings report. The internal organizational context graph Everest enabled two services—grocery delivery and airport pickup—to launch quickly, with the latter taking about six weeks to develop. Airbnb adopts a multi-model strategy, deploying at least 10 custom models and selecting models for different scenarios along the Pareto frontier of cost, performance, and latency: the strongest frontier models for coding, and smaller, faster post-trained models for search. The internal agent AirChat is already connected to MCP organizational context, and the team is advancing event-triggered asynchronous agents to automatically handle on-call.
English translation is being prepared.
ai-vertical
lab
English translation is being prepared.
Ai2 releases MolmoMotion: by inputting video frames, 3D points on objects, and action descriptions, it can predict the trajectories of these points in 3D space over the next few seconds, outperforming existing prediction methods. It also open-sources the MolmoMotion-1M dataset (derived from 1.16M videos) and the PointMotionBench benchmark (2.7K video clips), and releases the model weights. In simulation, a control policy based on MolmoMotion achieved a 76.3% success rate on pick-and-place tasks, higher than the 56.0% achieved by one based on Molmo 2.
English translation is being prepared.
tech-media
English translation is being prepared.
The Verge reporter Josh Dzieza recounted on a podcast his months-long investigation into Kevin O'Leary's Utah data center project Stratos (also known as Wonder Valley). The project is planned to cover 40,000 acres and supply 9 gigawatts of power, more than twice the average electricity consumption of the entire state of Utah. It bypassed local land approval through the state-level military installation development agency MIDA, and Box Elder County only learned of it one month before the vote. Local residents quickly organized opposition, causing two local commissioners and state Senate President J. Stuart Adams to lose their seats, and the project is currently paused due to referendum litigation. Dzieza believes that the core of data center development is speed, and local residents are never participants in this chain.
English translation is being prepared.
English translation is being prepared.
The US "One Big Beautiful Bill Act" expands the Opportunity Zone program; starting January 1 next year, more than 100 data center projects under construction in rural areas may newly qualify for corporate tax breaks. Research by the Searchlight Institute shows that Pew data indicates about 67% of planned data centers are sited in rural areas, while currently only 13% of operating data centers are in rural areas. Microsoft, Meta, and Amazon have all denied using the program, and Senator Josh Hawley has introduced legislation to eliminate Opportunity Zone funding for data centers.
English translation is being prepared.
a6-editorial
English translation is being prepared.
Cerebras had an AI agent run 71 experiments across two task types—training optimization and model compression—and found that when the loop scope is tightened, it reliably produces real conclusions, while once the guardrails are loosened, the agent drifts from the goal within hours. In experiment one, codex-autoresearch-harness was used to wrap Codex in a bash loop; GPT-5.4 and Codex-Spark each independently converged on a learning-rate warmdown schedule within 6 hours, but GPT-5.4 had a proposal acceptance rate of 67% versus only 17% for Spark, showing that proposal quality determines total cost more than single-inference speed. In experiment two, REAP plus INT4 quantization compressed a 717 GB model to 92 GB, with dynamic expert swap-in/swap-out taking 0.151 seconds, but after 12 hours of unattended operation the agent changed to asking a different question on its own; the authors corrected this by clearing context, isolating directories, and adding more frequent validation checkpoints.
English translation is being prepared.
English translation is being prepared.
The Cerebras blog points out that the key to looping agents is not autonomy but verifiability, and uses Gemma 4 running a visual feedback loop on Cerebras as an example: each iteration generates a new STEP file in about 1.2 seconds, at a speed of about 1,500 tokens/second, and after the loop autonomously rewrites the prompt about 5 times, it completes the CAD instruction to clone an object from a photo. The article also warns that the loop can spin idle when the goal is unclear, or be exploited by the model when the prompt is vague.
English translation is being prepared.
English translation is being prepared.
Cerebras published a post introducing the three GPT-5.6 models Sol, Terra, and Luna in the Codex subscription; the three are trained and deployed independently and come with reasoning-effort levels, used to trade off among speed, cost, and intelligence. On price, Terra is half of Sol, and Luna is one-fifth of Sol, for both short and long contexts. The article suggests starting most tasks with Luna and upgrading to Terra or Sol when progress stalls; Sol can reach 750 tokens/s on Cerebras, and it also explains how to use the five reasoning levels in Codex, from Light to Ultra.
English translation is being prepared.
chinese
English translation is being prepared.
English translation is being prepared.
Apple's stock bucked the trend on Monday, rising more than 1% to a record high of $336.91, with a market value of about $4.93 trillion, overtaking Nvidia ($4.78 trillion), which plunged nearly 5% that day, and reclaiming the title of the world's most valuable listed company. Apple has outperformed the Nasdaq 100 by 23 percentage points so far in July, its largest monthly excess gain since 2005, as investors view its restrained AI spending as an advantage. Apple pays Google only about $1 billion a year in licensing fees to use a customized Gemini model for Siri upgrades.