The Archive

The Alibaba logo is displayed outside its headquarters in Hangzhou, Zhejiang Province, China. | Image: NurPhoto via Getty Images Chinese tech giant Alibaba released what it says is its largest and "most capable AI model to date," claiming performance rivaling the best systems from US frontier labs Anthropic and OpenAI, as well as domestic rivals like Moonshot AI's Kimi K3. Alibaba said it was making the model, Qwen3.8-Max, widely available to users in a blog post published on Monday. The release had been expected after the company previewed the model last month, when it claimed it was "second...

June emerged from stealth today with a $20 million pre-seed round to make AI adoption simpler.

Meta's earnings miss and delayed AI product roadmap raise questions about execution timeline and capital allocation strategy.

MIT Technology Review Explains: Let our writers untangle the complex, messy world of technology to help you understand what’s coming next. You can read more from the series here. When two OpenAI models hacked into the website Hugging Face in July, they weren’t trying to make money or commit sabotage—they were just looking for answers…

OpenAI's GPT-Live enables low-latency continuous voice interaction via turnless speech model architecture, shipped in six months.

condense-json 1.1 adds support for non-string replacement values and object-based merge operations with round-trip testing.

Circles deploys OpenAI API and Codex for telecom personalization, achieving 22% ARPU lift and 9% churn reduction.

Simon Willison releases condense-json 1.0, a utility library for compressing JSON by replacing repeated strings with references.

On the latest episode of Equity, we discuss why Sam Altman has calling on the industry to "pace the rate of AI development."

Edward “Bud” Cole speaks in Japan in 2023. | Image: Jun Sato/WireImage Fender CEO Edward "Bud" Cole gave an interview to T3 in May celebrating the 75th anniversary of the Telecaster with comments on AI and music that initially flew under the radar. But it has started making the rounds recently, pouring more fuel on an already raging fire of bad PR following the company pissing off basically the entire guitar-playing community by sending cease-and-desist letters to builders, claiming copyright of the Stratocaster body shape. Some influential guitar YouTubers have even said they're done buying ...

Illustrators have spent years sounding the alarm about generative artificial intelligence startups training their models on artists' work without permission. They've pointed out how the practice is tantamount to theft, and in response, many gen AI boosters have argued that it's necessary for the technology's evolution. This has led to contentious legal battles, but it's also given rise to a new wave of AI startups like Pippa that market their services as being more ethical than the competition. Like most companies selling access to text-to-video models, Pippa's main product is short bursts of...

Role-decoupled attention residuals decouple matching and content retrieval in Transformer depth-routing, improving architectural efficiency.

Logit-origin centering enables singleton test-time adaptation for tabular models without batch dependencies.

Analysis of post-bandit inference bias under UCB1 and index algorithms via effective exploration rate characterization.

CompressAgent benchmark evaluates reliability of compressed agent control contexts across Qwen models and task families.

MCR-GRPO assigns marginal box contributions in multimodal RL for structured visual perception tasks without response-level broadcast.

Wix Helpmate deploys deterministic executability gating to filter skill selection in LLM agents by account state feasibility.

FactorJEPA applies factorized JEPA world models to chaotic Global South urban environments with high occlusion and agent heterogeneity.

DeBERTa-Sentinel uses disentangled attention for robust AI-generated text detection resistant to paraphrasing and model-diversity attacks.

Decoy images amplify caption-mediated defenses (ECSO) against encoded jailbreaks on VLMs, reducing attack success up to 73 percentage points.

Temporal replay framework evaluates enterprise agents against dynamic data across multiple moments within an episode, not just final state.

ML predictors for processor microarchitecture ranking fail on individual program phases despite strong aggregate performance.

LB-TrajRep proposes lower-bound representations for trajectory similarity without deep embeddings.

WAM-Diff2 distills autoregressive VLAs into efficient diffusion policies for low-latency autonomous driving.

Opt.Gear foundation model (1M–1B params, 64K context) uses hybrid conv-attention for 4.9× faster on-device inference.

CallScreenBench evaluates on-device LMs as phone secretaries with adversarial callers and no oracle ground truth.

Theoretical analysis of benign overfitting in linear regression with structured deterministic training data.

VLAGuard framework defends VLA robots against physical adversarial patches via attention-protective fine-tuning.

Stress-Relief Annealing optimizes automated warehouse layouts without simulation using polynomial-time methods.

Caliber defends against model extraction by adding calibrated Gaussian noise to logits with provable query costs.