There were 1 billion commits in 2025. Now, it's 275 million per week, on pace for 14 billion this year if growth remains linear (spoiler: it won't.)

benchmarks sourced from publicly available material carry contamination risk, where training-data exposure can silently inflate scores.

Meta founder and CEO Mark Zuckerberg described superintelligence in a blog post last year

Wedbush Securities analyst Dan Ives said in a report on Thursday.

The layoffs will start on May 20, the company confirmed.

Meta plans to lay off roughly 8,000 employees, or 10% of its workforce

Drug manufacturers pay pharmacy benefit managers rebates above 50% of list price for formulary access.

Google Cloud launched a parallel $750m fund to pay McKinsey, Accenture, and Deloitte to train engineers and co-fund client AI projects.

Structure: $500M OpenAI equity plus $4B from TPG, Bain, Advent, Brookfield, and Goanna form a $10B LLC.

OpenAI pledged $1.5B to a joint venture called DeployCo, guaranteeing private-equity partners a 17% annual return floor over five years.

Amazon is investing $5 billion in Anthropic today, with up to an additional $20 billion in the future

run-rate revenue has now surpassed $30 billion, up from approximately $9 billion at the end of 2025

committing more than $100 billion over the next ten years to AWS technologies

over one million Trainium2 chips to train and serve Claude

over 100,000 customers now run Claude on Amazon Bedrock

up to 5 gigawatts (GW) of capacity for training and deploying Claude

over one million Trainium2 chips to train and serve Claude

run-rate revenue has now surpassed $30 billion, up from approximately $9 billion at the end of 2025

Amazon is investing $5 billion in Anthropic today, with up to an additional $20 billion in the future

more than $100 billion over the next ten years to AWS technologies

over 100,000 customers now run Claude on Amazon Bedrock

up to 5 gigawatts (GW) of capacity for training and deploying Claude

For Anthropic, more usage across diverse tasks means more data, which produces a smarter model—just as more queries improved Google search.

For Anthropic, more usage across diverse tasks means more data, which produces a smarter model—just as more queries improved Google search.

Parameters are estimated by unweighted least squares. Time t is measured in years since the first observation in each dataset.

We pre-selected the 6-month horizon as our primary metric, balancing genuine forecasting distance against the limited date range of our data.

The minimum training cutoffs are: ECI (June 2024), METR Time Horizon (January 2024), Combined Math (September 2024), and WeirdML V2 (January 2025).

Our fourth metric, an index constructed from WeirdML V2 results, showed no sign of acceleration. A single global linear trend fit the data best.

We use four AI capability metrics: ECI (Epoch Capabilities Index), METR 50% Time Horizon, Combined Math Index, and WeirdML V2 Index.

Reasoning models show both a one-off jump in performance and a roughly 2-3x faster trend compared to non-reasoning models.

Three of the four metrics (ECI, log METR 50% time horizon, and a math-focused index we constructed from several math benchmarks) show strong evidence that progress has sped up relative to a global linear trend fit to data from 2023 onward.

The three metrics where we find acceleration are concentrated in programming and mathematics. These are areas that labs have explicitly targeted for improvement

Our fourth metric, an index constructed from WeirdML V2 results, showed no sign of acceleration. A single global linear trend fit the data best.

The best-performing model across these three metrics was a pair of independent linear trends: one for reasoning models and one for non-reasoning models.

Reasoning models show both a one-off jump in performance and a roughly 2-3x faster trend compared to non-reasoning models.

Three of four metrics show strong evidence of acceleration, driven by reasoning models.

The three metrics where we find acceleration are concentrated in programming and mathematics.

We select the median-difficulty question from the set with maximum model coverage and standardize it to 0.

We work with the natural logarithm of the time horizon, which puts it on an approximately linear scale.

The minimum training cutoffs are: ECI (June 2024), METR Time Horizon (January 2024), Combined Math (September 2024), and WeirdML V2 (January 2025).

Reasoning models show both a one-off jump in performance and a roughly 2-3x faster trend compared to non-reasoning models.

Three of four metrics show strong evidence of acceleration, driven by reasoning models.

Unfortunately, the attacker got further access through their enumeration.

Vercel stores all customer environment variables fully encrypted at rest. We have numerous defense-in-depth mechanisms to protect core systems and customer data.

SWE-chat is a living dataset; our collection pipeline automatically and continually discovers and processes sessions from public repositories

frontier AI models are not too big because the technology is complex and too big because the training data is garbage

The modern data stack has undergone a decade+ transition from disparate data sources to consolidated data and cleaned definitions (which is good), but even then the consolidation is never perfect and a lot of messiness is introduced.

To overcome this blocker, a team member hard codes the exact revenue and timeframe definitions. The data agent continues chugging along but quickly runs into challenge #2 – where are the right data sources? Which ones are the right sources of truth?

Over the past year, the market has realized that data and analytics agents are essentially useless without the right context – they aren't able to tease apart vague questions, decipher business definitions, and reason across disparate data effectively.

包括 47 种数据：agents、skills、hooks、MCP 配置、会话历史、自定义规则……

多年积累的对话、定制 Agent、项目记忆、MCP 配置、Skill 库——一次风控就可能全部失联。

They provide access to more than 50 public multi-omics databases, literature sources, and biology tools, and offer a flexible starting point for common repeatable workflows.

Some privacy related extensions may cause issues on x.com.

Some privacy related extensions may cause issues on x.com.

There were 1 billion commits in 2025. Now, it's 275 million per week, on pace for 14 billion this year if growth remains linear

We present a comprehensive adoption snapshot of the leading open language models and who is building them, focusing on the ~1.5K mainline open models

Sage sends URLs and package hashes to Gen Digital reputation APIs. File content, commands, and source code stay local.

Academic publishers, documentary archives, game studios, and companies sitting on years of enterprise data have all been courted for the seeds of intelligence needed to train the next generation of models.

Mercor, which provides data to AI labs for training, became one of the fastest-growing companies in history before losing four terabytes of data to hackers last week.

A small model trained on fewer than 2,000 examples from real lawyers, bankers, and consultants recently beat all but the best frontier models on corporate legal work, at a fraction of the price.

Reddit, Shutterstock, and News Corp are making hundreds of millions a year licensing their high-quality data to companies training AI, and those contracts are growing about 20 percent annually, according to their quarterly filings.

Our Chip Ownership data does not capture all global chip ownership, and has weaker coverage prior to 2023.

As slop takes over the Internet, labs may struggle to obtain high-quality corpuses for training models.

Based on our analysis, **29% of the Fortune 500 and ~19% of the Global 2000**are live, paying customers of a leading AI startup.

Support teams are high volume and high turnover, and thus need to train new reps in a fast and standardized way. To do so, they have clearly articulated standard operating procedures (SOPs) that guide the work of each rep. These SOPs create clear rules and guidelines that AI agents can model themselves off of.

Closed harnesses behind proprietary APIs force yielding control of agent memory to third parties.

Within a few months, they have more than a dozen production enterprise deployments & are processing over a billion events per hour.

Maine advances first statewide moratorium blocking data centers requiring over 20 megawatts

the most interesting detail here is how SkillClaw clusters cross-user trajectories into referenced skills and then uses the evolver to translate those patterns into concrete updates.

We test for a trend over time by fitting a weighted linear model to the log-odds of usage. Under this specification, Claude is the only AI service in the survey to show a statistically significant upward trend over this period

The ChatGPT for Excel add-in operates separately from your ChatGPT chat history. Conversations and data in Excel aren't shared with your ChatGPT chats, and activity doesn't sync between experiences at this time.

By default, data shared with ChatGPT isn't used to improve our models for ChatGPT Business, ChatGPT Enterprise, ChatGPT Edu, and ChatGPT for Teachers.

we collaborated with over 1,000 physicians to curate training data that enables more factual and comprehensive responses.

The model reportedly scored 93.9% on SWE-bench Verified and 77.8% on SWE-bench Pro, but its strongest signal came from real-world results, including uncovering a 27-year-old flaw in OpenBSD, a 16-year-old vulnerability in FFmpeg, and autonomously chaining Linux kernel exploits without human input.

We need, like, a Manhattan Project to collect this

We need, like, a Manhattan Project to collect this... Fields that are not exposed now will become exposed in the future, so you just want to track these statistics across the entire economy.

A learning system can continuously incorporate real-world data in a way that numerical solvers fundamentally cannot, capturing and compounding the knowledge that is currently trapped out there in the real world.

inappropriately change or overwrite JSON files compared to Markdown files

按时间记录不完全合理，还是应该按任务记录。

βテスト期間中のご利用は無料です。

American hyperscalers are driving a data center buildout that's larger than the Manhattan Project and Apollo Program at their peaks.

MiniMax may have been able to get 100 billion tokens of data from interactions with Claude.

A three-stage progressive training strategy -- large-scale pre-training, hard sample fine-tuning, and GRPO alignment -- sequentially exploits these data at different quality tiers.

SOTA models of different architectures and parameter scales exhibit highly consistent failure patterns on the same set of hard samples, suggesting that the performance bottleneck stems from shared deficiencies in training data rather than architecture itself.

Without any architectural modification, MinerU2.5-Pro achieves 95.69 on OmniDocBench v1.6, improving over the same-architecture baseline by 2.71 points and surpassing all existing methods including models with over 200× more parameters.

Current document parsing methods compete primarily on model architecture innovation, while systematic engineering of training data remains underexplored.

introducing a commercial text and data mining exception for AI training would expand the AI sector in the country.

most existing large language model agent systems face severe limitations in data-intensive settings, including context saturation, cascading error propagation, and high end-to-end latency

We introduce Iterative Reward Calibration, a methodology for designing per-turn rewards using empirical discriminative analysis of rollout data

If we knew that every image uploaded was a beautiful model shot, segmentation would be far easier, but because of the nature of user-uploaded content, we need the best possible segmentation.

Urgent treatment for neoplasm consists of (1) cautious use of intravenous diuretics and (2) mediastinal irradiation, starting within 24 hours, with a treatment plan designed to give a high daily dose of radiation but a short total course of therapy to rapidly shrink the local tumor. Intensive radiation therapy combined with chemotherapy will palliate the process in up to 90% of patients. In patients with a subacute presentation, radiation therapy alone usually suffices. Chemotherapy is added if lymphoma or small-cell carcinoma is diagnosed

Interviews were video and audio recorded. We transcribed the audio using OpenAI's Whisper automatic speech recognition system and anonymized the transcript before analysis. We analyzed the interview data using thematic analysis [1]. First, two members of the research team independently coded four (25% of collected data) randomly chosen participant data to generate low-level codes. The inter-coder reliability between the coders was 0.88 using Krippendorff's alpha [37]. The two coders then met together to cross-check, resolve coding conflicts, and consolidate the codes into a codebook across two sessions. Using the codebook, the two coders analyzed six randomly selected participant data each. The research team then met, discussed the analysis outcomes, and finalized themes over three sessions.

We conducted a qualitative analysis of user study transcripts and survey responses using a Grounded Theory approach [8]. First, the lead researcher collected a list of participants' behaviors, approaches, reflections on their experience, and feedback about the interface. The researcher then systematically coded this data, revisiting the data multiples times and refining the codes to ensure consistency and coherence. Through this process, high-level themes were identified and organized using affinity diagramming. Once the thematic structure was finalized, the researcher gathered supporting evidence for each theme and synthesized the findings, which were reviewed by the research team to ensure agreement on the results.

Activity log data, which revealed how participants actually used the interface, echoed the above findings. According to the log data, participants spent most of their reading time (66.31%) with vertical alignment on the second element in structure pairs, followed by alignment on the first element (29.19%), and left-justified alignment (5.13%). Highlighting usage showed a similar preference: 91.13% of time with all chunks highlighted, 8.25% with partial highlighting, and minimal time (0.63%) without highlights.

In this section, we present findings on how AbstractExplorer supports comparative close reading at scale by integrating quantitative survey responses and log data with qualitative analysis of transcripts and open-ended responses. The qualitative analysis process is described in detail in Appendix H.

Throughout the two tasks, we also collected detailed interaction logs including counts of user-defined aspects created, duration of highlighting usage, and time allocation across the three possible alignment options.

Both gaze data and the semi-structured interviews revealed that lower NFC participants were more willing to be guided by the three features and took advantage of them consciously.

Using a two-tailed Mann-Whitney U Test, we found that participants who reported their lowest perceived cognitive load when all three features were enabled had significantly lower NFC than participants who reported their lowest cognitive load level when skimming with no features enabled—in the baseline interface (p=0.03).

The raw NASA-TLX score is the sum of all 6 NASA-TLX questions after reversing the appropriate questions.

To compute a participant's NFC score, we averaged their response to the six questions, each ranging from 1 to 7, after reversing the appropriate questions.

For simplicity of analysis, we denote participants with NFC scores above the overall participants' median NFC of 5.42 (IQR = 0.583) as higher NFC, and lower NFC otherwise.

To contrast participants' gaze patterns in each condition, we used a Tobii Pro Spark eye-tracker placed below the desktop monitor used by all subjects; Tobii Pro Lab software recorded each participant's gaze over time in each condition.

We collected 80 sentences from our abstracts dataset labeled by our system as "Methodology/Contribution." Participants viewed the same 80 sentences in each condition—often with a different subset of sentences initially visible due to ordering changes—but only had two minutes to look at them in each condition.

After obtaining an expanded set of high-level chunk labels, we assign them to each of the sentence chunks by using LLMs in a multiclass classification few-shot learning task, with the initial labels and assignment as examples (see prompt used in Appendix D.3).

Then, we segment sentences within each aspect into grammarpreserving chunks (see prompt used in Appendix D.2). This results in grammatically coherent chunks that are the basis of structure patterns. After identifying chunk boundaries, we again prompt an LLM to generate labels for chunks in a human-in-the-loop approach: starting from an initial set of labels for chunk roles, when a new label is generated, a researcher from the research team examines the new label and merges it with existing labels if appropriate, controlling for the total number of labels.

We process this data in a three-stage pipeline (Figure 6). In the first stage, Sentence Segmentation and Categorization, abstracts are split into individual sentences using the NLTK package, and each sentence is classified into one of the five pre-defined aspects as listed in Section 4.1.1. Classification is performed by prompting an LLM (see prompt used in Appendix D.1) with the sentence and its full abstract.

After the interviews, we analyzed the data using the process described in Appendix B

To analyze the annotation efficiency, we first conducted a Kruskal-Wallis rank sum test [39] to determine if there were statistically significant differences in annotation time across the three conditions, because our data violated the homogeneity of variances assumption, making non-parametric methods more appropriate.

Figure 3: Decrease in Perceived Fiscal Benefits of Microcredentials by Year

Notably, traditional mindsets and legacysystems are seen as far greater barriers in 2025 (61%) than in 2021 (5%), high-lighting a growing tension between innovation and institutional resistance tochange.

Comprehensive Learner Record Standard Transcript Guide

23

If you value your data I suggest not trusting any filesystem or media, consider them all equally fallible.

Media files are not directly downloaded in overall syncing to save bandwidth. Instead, when that file is requested, it is streamed to your device from the backup node or your devices on the network. For example, if you have a 4K Video, it will be streamed from the backup node or P2P devices to your device. So when you open an object with an image, it downloads. When you press play on video & audio, it begins to download. After that, this file will be stored in the application cache.

Benioff had recently told Business Insider that he's drafting the company's annual strategic document with data foundations—not AI models—as the top priority, explicitly citing concerns about "hallucinations" without proper data context.

our world and data does they do have some legitimate research because that's what think tanks do. They launder illegitimate research with legitimate research. uh and their tactic primarily is to uh set the scope of what they are commenting on or researching uh that it you know it puts forward the kind of results that they want uh that aligns with their ideology.

Micro-Credentials and Digital Badges: An Exploration of Definitions and Implications in Higher Education and Workforce

Systolic blood pressure.

3.1 Physiological response of viewing different landscape types

How important for being looked up to or having high status in your school is...

How often do you feel...

How much competition for grades is there�

During an average school week, about how many times�

NeverLess than once a week1-2 times a week3-5 times a week6-9 times a week10-19 times a week20 or more

I feel I am a person of worth, on an equal plane with others

I will not link to the photo here. Allow me my parental illusions of protection. I reprint it here on social media only in doctored form. I didn’t know what to use to cover her eyes.

During each call, Stewart said, Amazon officials have not been helpful."They wanted to do background checks on all my firefighters; I wouldn't let them," he said. "And we've struggled to gain access to emergencies. They'll stop us at the gate, and our medic units have been delayed. They're denying us access to patients.

Recently, OpenAI has shared something. In a blog post, CEO Sam Altman said that the average query uses about 0.34 watt hours of energy.

When you open this in two browsers and refresh a few times, one browser after the other, you’ll see the count go up and up (when looking at the page source), proving that the state is shared between both browsers (well, not really, it’s shared on the server, and used by both users). This will have serious consequences if you go this route: if user A is logged in and you’d write the user object to the shared state, and user B is not logged in, they’d still see a flash of user A’s username appear in the navigation bar, until the shared state is overwritten by the undefined user object.

export const state: State = $state({ user: undefined });

But sadly this introduces shared state on the server (when we use SSR), and this is a big problem since we’re now leaking data between different users.

One pattern that I love to use in my SvelteKit projects is returning writable stores from the layout’s load function. This makes it possible to fetch data from the server (for example the user object for the logged in user), and then you make this object available as a writable reactive store throughout the whole application. So when the user updates their username or avatar, you do the PUT request to the server and you get the updated user object back from the server as the response, you can simply update the $user writable store value and every place in your app where you show the user object gets updated immediately.

risk of accidentally exposing one user’s data to another

As with the previous example, this puts one user’s information in a place that is shared by all users.

But what if you want to update this user instance? For example on your website you have a form where the user can change their name, username, or avatar. When the form is submitted this gets stored on the server, but the site still shows the old user information, for example it still shows the old avatar of the user in the top menu. The user variable isn’t writable, so how do you overwrite this?

Stateless vs. Stateful Preprocessing: Most PyTorch transforms are stateless (e.g., RandomHorizontalFlip) or configured with fixed parameters (e.g., Normalize with pre-defined mean/std). If you need to compute statistics from your data (like the mean and standard deviation for normalization), you typically do this once offline and then hardcode these values into the Normalize transform. This contrasts with Keras's Normalization layer, which has an adapt() method to compute these statistics online from a batch of data.

Preprocessing challenges The following are the primary challenges of implementing data preprocessing: Training-serving skew. Training-serving skew refers to a difference between effectiveness (predictive performance) during training and during serving. This skew can be caused by a discrepancy between how you handle data in the training and the serving pipelines. For example, if your model is trained on a logarithmically transformed feature, but it's presented with the raw feature during serving, the prediction output might not be accurate. If the transformations become part of the model itself, it can be straightforward to handle instance-level transformations, as described earlier in Option C: TensorFlow. In that case, the model serving interface (the serving_fn function) expects raw data, while the model internally transforms this data before computing the output. The transformations are the same as those that were applied on the raw training and prediction data points. Full-pass transformations. You can't implement full-pass transformations such as scaling and normalization transformations in your TensorFlow model. In full-pass transformations, some statistics (for example, max and min values to scale numeric features) must be computed on the training data beforehand, as described in Option B: Dataflow. The values then have to be stored somewhere to be used during model serving for prediction to transform the new raw data points as instance-level transformations, which avoids training-serving skew. You can use the TensorFlow Transform (tf.Transform) library to directly embed the statistics in your TensorFlow model. This approach is explained later in How tf.Transform works. Preparing the data up front for better training efficiency. Implementing instance-level transformations as part of the model can degrade the efficiency of the training process. This degradation occurs because the same transformations are repeatedly applied to the same training data on each epoch. Imagine that you have raw training data with 1,000 features, and you apply a mix of instance-level transformations to generate 10,000 features. If you implement these transformations as part of your model, and if you then feed the model the raw training data, these 10,000 operations are applied N times on each instance, where N is the number of epochs. In addition, if you're using accelerators (GPUs or TPUs), they sit idle while the CPU performs those transformations, which isn't an efficient use of your costly accelerators. Ideally, the training data is transformed before training, using the technique described under Option B: Dataflow, where the 10,000 transformation operations are applied only once on each training instance. The transformed training data is then presented to the model. No further transformations are applied, and the accelerators are busy all of the time. In addition, using Dataflow helps you to preprocess large amounts of data at scale, using a fully managed service. Preparing the training data up front can improve training efficiency. However, implementing the transformation logic outside of the model (the approaches described in Option A: BigQuery or Option B: Dataflow) doesn't resolve the issue of training-serving skew. Unless you store the engineered feature in the feature store to be used for both training and prediction, the transformation logic must be implemented somewhere to be applied on new data points coming for prediction, because the model interface expects transformed data. The TensorFlow Transform (tf.Transform) library can help you to address this issue, as described in the following section.

You preprocess the raw training data using the transformation implemented in the tf.Transform Apache Beam APIs, and run it at scale on Dataflow. The preprocessing occurs in the following phases: Analyze phase: During the analyze phase, the required statistics (like means, variances, and quantiles) for stateful transformations are computed on the training data with full-pass operations. This phase produces a set of transformation artifacts, including the transform_fn graph. The transform_fn graph is a TensorFlow graph that has the transformation logic as instance-level operations. It includes the statistics computed in the analyze phase as constants. Transform phase: During the transform phase, the transform_fn graph is applied to the raw training data, where the computed statistics are used to process the data records (for example, to scale numerical columns) in an instance-level fashion.

Stackable credentials are also critical to the “Some College, No Credential” (SCNC) market, which reached a total of 36.8 million under the age of 65 in the U.S., up 2.9% from the previous year. Recent research from UPCEA and StraighterLine found that 76% of SCNC adults said being able to earn alternative or microcredentials that could stack toward a degree would increase or greatly increase their interest in completing their degree

Data Frame

fighter1 data