Renee Li settled in New York after living in Asia, Europe, and South America. Renee Li Analytics engineer Renee Li moved to New York City about three years ago. Li said it's the most expensive city she's ever lived in, but there are ways to make it affordable. She uses apps, credit card points, and other hacks to get deals in the city. This as-told-to essay is based on conversations with Renee Li, a 29-year-old analytics engineer in New York. It has been edited for length and clarity. I've been in New York for about three years. Before that, I was fully nomadic in Asia, South America, and Europe, living out of a carry-on. Most of my friends were in New York, and over time, I found myself spending a quarter to half the year here. Eventually, I got a remote analytics engineer job at a fintech company based in the city, and it gave me another reason to settle down here. I wanted a place to call home. I make $250,000 a year as an analytics engineer, get equity, and invest in real estate. Plus, I play poker, which can go either way financially. Sometimes it's income, sometimes it's a loss. Are you a New Yorker working to afford the city? Reach out to this reporter at mros@insider.com New York is the most expensive place I've lived, by far. As much as I complain about the fees and late subways, New York is still hard to beat. I pay $3,500 for a one-bedroom in Nolita, near SoHo in Manhattan I moved near SoHo in a 400-square-foot apartment for the convenience of being steps from multiple subway lines, grocery stores, gyms, restaurants, and cafés. Food and dining are another major expense. I spend about $650 a month on food, mostly eating out, with 20% on groceries. Shopping is usually around $600 to $800 a month, including clothes, skincare, furniture, and household items. Transportation is about $300 a month, including rideshares, public transit, and biking. Fitness is another major category. Pilates is $360 a month, and I also pay $60 for ClassPass. In travel-heavy months, experiences and activities like tours and shopping become a major expense — I travel hack and often redeem points for flights. Li prefers living in the city to be near subways, grocery stores, restaurants, and shops. Renee Li Gas and water are included in my rent. My internet, electricity, and phone bills don't amount to much. I pay about $200 a month for healthcare after my employer's contribution. I refuse to buy anything full price in New York. I use apps, credit card points, and other hacks. One of my biggest financial levers is that I graduated on a full-ride from my dream college, Rice University. Starting my career debt-free meant I didn't have the mental burden of constantly thinking about student loans. To make the most of living in New York, I've figured out hacks for almost every spending category — mostly cash back or points. For dining, I use apps like Seated and InKind, linking my credit card to earn cash back. For travel, I'm big on credit card points and travel hacking. I book most of my flights with points, which saves a lot. I try not to spend too much in the city beyond my day-to-day expenses. I refuse to buy new clothes here. I'd much rather spend that $100 at a quality secondhand store or purchase clothing in Asia or Europe. Li sometimes travels to Asia for shopping, healthcare, and beauty. Renee Li It's the same with skincare and hair. I'll stock up on skincare in Korea or other parts of Asia, and schedule hair and beauty appointments while overseas. In New York, those services can easily cost $300 to $600. There are many ways to make the city more affordable NYC is not as expensive as you think. It can be affordable to eat if you're not eating at the trendy restaurants. For example, I get lunch in Chinatown for around $8. The same goes for beauty and wellness. If you go farther out for haircuts, lashes, eyebrows, skincare, or massages, you can save a lot. In Chinatown, I can get a massage for around $50 an hour, compared to $300 in NoMad. Working for a New York company also helps because salaries tend to be higher, which helps offset costs. I don't feel much pressure to keep up appearances. I work in tech, which is pretty casual and laidback. One thing that still shocks me about New York is the fees It's not just gratuity, but service fees, city tax, and state tax. I'd rather pay a higher upfront price than see one price on the menu and then have gratuity, tax, and other misc fees added at checkout. Housing fees are frustrating, too. If you're applying for a sought-after building, there might be administrative fees. If you don't pay, other people are lined up for the unit. I don't think I'll stay in New York forever. If I ever had kids, my salary could feel stretched. Much of that income disappears once you add childcare or private school. Still, I wouldn't want to live in the suburbs because I hate driving. I also don't want to deal with buying a house right now. It's easy to feel behind in New York, but I don't feel financial pressure I grew up low-income with a single parent, and as the oldest child, I felt a lot of that burden. My childhood gives me a greater appreciation for things that were once luxuries and are now part of my day-to-day life. In the New York bubble, it's easy to feel behind. There are people with trust funds or making $750,000 in hedge funds. Once I zoom out to include the rest of the world, I realize I'm actually doing really well, especially compared with where I grew up. I've lived in DC and San Francisco, but New York feels unmatched for career opportunities and diversity of industries and people. It's an incredible place to be in your 20s if you can make it work. The biggest advantage of being in New York is the serendipity and opportunity. You can meet future partners, business partners, or best friends by being out and about. Do you want to share your cost of living with Business Insider? Email editor Manseen Logan at mlogan@businessinsider.com. Read the original article on Business Insider

This week, I covered the story of Casey Harrell—a man with ALS who is “the first power user” of a brain implant, according to the researchers who worked with him. Harrell is paralyzed and unable to speak coherently without the device. He has now spent almost three years using a brain-computer interface (BCI) that enables…

There are plenty of useful things a metric can reveal. There are even more it can obscure or corrupt. It took me well over a decade of tracking my own life in ever greater detail to fully appreciate this duality, which probably reveals something about both me and the nature of measurement. Like a lot…

PLUS: Save hours on meeting prep with Google Gemini

Australian enterprises remain dependent on foreign technology infrastructure, while major economies treat that dependency as a strategic liability. The post As the World Claims Tech Sovereignty, Where Does Australia Stand? appeared first on TechRepublic.

There's a commercial logic that cuts against the idea that ASML would risk its export license to arm a Chinese customer.

Discussion | Link

Discussion | Link

The Data Centralization Myth: Why Transformation Needs Governed Decentralization  Companies spend millions on warehouses, lakehouses, and … Digital Transformation Starts Where Decisions Happen, Not Where Data Is Stored Read more The post Digital Transformation Starts Where Decisions Happen, Not Where Data Is Stored  appeared first on .

Wall Street Journal: Sources detail how Google is using Nvidia's playbook to build an AI chip business, including providing $3.2B to fund a NY data center renting TPUs to Anthropic  —  Wielding its war chest to win data-center customers for its silicon, the world's second-biggest company is taking a page from No. 1

Discussion | Link

Discussion | Link

Hayden Field / The Verge: Barret Zoph leaves OpenAI again five months after rejoining in January; Zoph initially left OpenAI in 2024 to serve as Thinking Machines Lab co-founder and CTO  —  Five months after returning to OpenAI, Barret Zoph — the company's head of enterprise AI sales — has departed, The Verge has learned.

Bloomberg: Sources: about 200 companies accessing Mythos Preview through Project Glasswing have preserved their access despite a recent US government shutdown order  —  Some firms chosen early on by Anthropic PBC to test the Mythos AI model ahead of a wider release have preserved their access to a preview …

Jesse Coghlan / Cointelegraph: The CFTC permanently bans Celsius founder Alex Mashinsky from trading in markets it oversees as part of a settlement resolving its 2023 lawsuit against him  —  Cointelegraph is committed to independent, transparent journalism.  This news article is produced in accordance …

arXiv:2606.19364v1 Announce Type: new Abstract: The prefill stage of Large Language Model (LLM) inference is a growing contributor to cloud-scale energy cost. Many consumer-support and conversational prompts contain social scaffolding: politeness markers, apologetic preamble, repetition, and rapport-building language that is important for human communication but carries low marginal information for machine reasoning. We call this discrepancy the Social-Semantic Gap. We present SPSD (Sentiment Preserving Semantic Distillation), an edge-based pipeline that compresses user prompts using a 4-bit quantised Small Language Model before transmission to a cloud-deployed LLM. Evaluation on a 248-prompt corpus using Gemma-2-2B-Instruct (Q4_K_M) as the SLM and Llama-3.1-8B-Instruct as the cloud evaluation model yields a mean input token saving of 99.9 tokens per distilled call, with all 146 distilled calls yielding positive savings. Response quality, assessed by blind LLM-as-judge scoring across 121 pairs, is non-inferior to the raw path within a pre-specified 1-point margin on a 15-point rubric; the judge awarded 43 percent ties, 28 percent distilled wins, and 29 percent raw wins. Cosine similarity is mixed: mean 0.682, median 0.712, with 54.1 percent of pairs above the 0.70 reference threshold. Safety-critical domains are conservatively routed to passthrough via rule-based gates. Per-call net energy saving is estimated at 70-270 uWh under stated assumptions. SPSD shows that on-device prompt distillation can reduce cloud LLM input-token cost while preserving response quality within a practical non-inferiority margin.

arXiv:2606.19584v1 Announce Type: new Abstract: Vision foundation models are typically trained as static feature extractors, placing the burden of task adaptation onto large downstream models. We propose an alternative paradigm: instead of solely feeding visual features into language models, we use language itself to dynamically guide the vision encoder. Our method, Language-Instructed Vision Embeddings (LIVE), leverages language as high-level guidance to produce task-centric embeddings at inference time, removing the need for task-specific retraining. This enables the encoder to focus on contextually relevant aspects of the input, yielding more controllable and generalizable representations. Empirically, LIVE reduces visual hallucinations (+34 points on MMVP), surpasses vision-language models with orders of magnitude more parameters on visual question answering, and generalizes to unseen instructions and tasks -- offering a direct path toward adaptive, instruction-driven visual intelligence.

arXiv:2606.19350v1 Announce Type: new Abstract: Large language models (LLMs) excel at multi-step reasoning but incur substantial inference cost. We introduce Causal Attribution Pruning (CAP), a training-free method that identifies critical attention heads by measuring their causal impact on reasoning tasks and uses these head-level scores to guide fine-grained weight pruning. For each attention head, CAP estimates the expected performance degradation when the head is masked during forward passes on a small calibration set of reasoning problems. These causal scores are then converted into weight-level importance values for the corresponding projection matrices. Unlike magnitude-only or activation-based criteria, CAP's interventional measurement directly captures each head's functional contribution, yielding relative accuracy gains of up to 61% over Wanda on ARC-Challenge at 20% sparsity. We evaluate CAP on GSM8K, StrategyQA, and ARC-Challenge using Llama-3-8B-Instruct and Mistral-7B-Instruct at 10%, 20%, and 50% sparsity. At moderate sparsity (10-20%), CAP improves over Wanda in most model-benchmark configurations. with especially large gains on ARC-Challenge for Llama-3. Our results suggest that attention-head-level causal attribution can better preserve reasoning performance on downstream benchmarks than correlational pruning criteria at equivalent sparsity, while remaining limited by coarse MLP attribution at 50% sparsity.

arXiv:2606.19344v1 Announce Type: new Abstract: Large Language Models (LLMs) exhibit representational and syntactic biases that are difficult to evaluate due to the stochastic nature of text generation. Standard auditing methods rely on a single output inspection or static automated metrics. These approaches obscure the underlying probability distributions and fail to capture biases hidden in lower-probability generation branches. This paper introduces TreeTracer, a visual analytics tool designed to evaluate LLM bias through aggregated comparison. Using a systematic perturbation analysis pipeline, the tool replaces ontology-defined terms in each input prompt, aggregates hundreds of stochastic generations into a syntax-aligned hierarchical structure, and then performs classification-aware node merging with an auxiliary language model. The resulting structure is visualized through a custom Sankey diagram. By juxtaposing two ontology-driven trees, the workspace enables direct comparison between semantic contexts and supports systematic bias detection. Because any visualization reflects only a subset of the model's learned behavior, the system further applies contrastive inference to compute and directly display counterfactual token probabilities across contexts, reducing the risk of misinterpreting the presence of bias. We validate the workspace through case studies comparing an unaligned baseline model GPT-2 XL against the constitutionally aligned Apertus models. The visual aggregation successfully exposes hidden representational harms, such as counterfactual pronoun suppression and conversational marginalization of individuals. A preliminary user study confirms that the aggregated comparative interface reduces cognitive load and effectively supports analysts in detecting systemic biases.

arXiv:2606.19483v1 Announce Type: new Abstract: Vision Foundation Models (VFMs) with Vision Transformer (ViT) backbones, such as DINOv2, have become essential for downstream tasks like object recognition and semantic segmentation. The immense computational requirements of backbones often necessitate distillation into smaller architectures for edge deployment. Feature-based knowledge distillation (KD) often suffers from the teacher-student gap; the student struggles to imitate teacher's complex feature map due to its limited capacity. To mitigate this bottleneck, we propose LEAP: Layer-skipping Efficiency via Adaptive Progression, a training curriculum for ViT feature-based knowledge distillation. By utilizing the teacher's intermediate feature maps as a sequence of progressively more difficult targets, our curriculum allows the student to build a foundational representation before tackling higher-level abstractions. Our results demonstrate that this paradigm significantly accelerates convergence through adaptive difficulty selection across various student model sizes and dataset scales. With our curriculum, the LEAP-distilled ViT-S achieves 90.1% accuracy on ImageNet-100, a +12.24% improvement compared with baseline. On ImageNet-1K, LEAP achieves +3.84% and +7.75% improvement for the instance retrieval task on the Oxford and Paris datasets, respectively. Furthermore, the curriculum enables 25.1% savings in training FLOPs and 21% savings in training time on ImageNet-100 by implementing early-stopping for teacher inference during the initial stages of training. Code is available at https://github.com/KevinZ0217/LEAP