Bitstrips, Inc. was a Canadian media and technology company based in Toronto, founded in 2007 by Jacob Blackstock, David Kennedy, Shahan Panth, Dorian Baldwin, and Jesse Brown. The company created and offered a web application, Bitstrips.com, which allowed users to create comic strips using personalized avatars, and preset templates and poses. Brown and Blackstock explained that the service was meant to enable self-expression without the need to have artistic skills. Bitstrips was first presented in 2008 at South by Southwest in Austin, Texas, and the service later piloted and launched a version designed for use as educational software. The service achieved increasing prominence following the launch of versions for Facebook and mobile platforms. In 2014, Bitstrips launched a spin-off app known as Bitmoji, which allows users to create personalized stickers for use in instant messaging. In July 2016, Snapchat Inc. announced that it had acquired the company; the Bitstrips comic service was shut down, but Bitmoji remains operational, and has subsequently been given greater prominence within Snapchat's overall platform. == History == Bitstrips was co-developed by Toronto-based comic artist Jacob Blackstock and his high school friend, journalist Jesse Brown. The service was originally envisioned as a means to allow anyone to create their own comic strip without needing artistic skills. Brown explained that "it's so difficult and time-consuming to tell a story in comic book form, drawing the same characters again and again in these tiny little panels, and just the amount of craftsmanship required. And even if you can do it well, which I never could, it takes years to make a story." Brown stated that the service would be "groundwork for a whole new way to communicate", and went as far as describing the service as being a "YouTube for comics". Blackstock explained that the concept of Bitstrips was influenced by his own use of comics as a form of socialization; a student, Blackstock and his friends drew comics featuring each other and shared them during classes. He felt that Bitstrips was a "medium for self-expression", stating that "It's not just about you making the comics, but since you and your friends star in these comics, it's like you're the medium. The visual nature of comics just speaks so much louder than text." The service was publicly unveiled at South by Southwest in 2008. In 2009, the service introduced a version oriented towards the educational market, Bitstrips for Schools, which was initially piloted at a number of schools in Ontario. The service was praised by educators for being engaging to students, especially within language classes. Brown noted that students were using the service to create comics outside of class as well, stating that it was "so gratifying and shocking what people do with your tool to make their own stories in ways that you never would have anticipated. Some of them are just brilliant." In December 2012, Bitstrips launched a version for Facebook; by July 2013, Bitstrips had 10 million unique users on Facebook, having created over 50 million comics. In October 2013, Bitstrips launched a mobile app; in two months, Bitstrips became a top-downloaded app in 40 countries, and over 30 million avatars had been created with it. In November 2013, Bitstrips secured a round of funding from Horizons Ventures and Li Ka-shing. In October 2014, Bitstrips launched Bitmoji, a spin-off app that allows users to create stickers featuring Bitstrips characters in various templates. In July 2016, following unconfirmed reports earlier in the year, Snapchat Inc. announced that it had acquired Bitstrips. The company's staff continue to operate out of Toronto, but the original Bitstrips comic service was shut down in favour of focusing exclusively on Bitmoji, leaving many Bitstrips users to call for a reboot of the comic service.
Text-to-video model
A text-to-video model is a form of generative artificial intelligence that uses a natural language description as input to produce a video relevant to the input text. Advancements during the 2020s in the generation of high-quality, text-conditioned videos have largely been driven by the development of video diffusion models. == Models == There are different models, including open source models. Chinese-language input CogVideo is the earliest text-to-video model "of 9.4 billion parameters" to be developed, with its demo version of open source codes first presented on GitHub in 2022. That year, Meta Platforms released a partial text-to-video model called "Make-A-Video", and Google's Brain (later Google DeepMind) introduced Imagen Video, a text-to-video model with 3D U-Net. === 2023 === In February 2023, Runway released Gen-1 and Gen-2, among the first commercially available text-to-video and video-to-video models accessible to the public through a web interface. Gen-1, initially released as a video-to-video model, allowed users to transform existing video footage using text or image prompts. Gen-2, introduced in March 2023 and made publicly available in June 2023, added text-to-video capabilities, enabling users to generate videos from text prompts alone. In March 2023, a research paper titled "VideoFusion: Decomposed Diffusion Models for High-Quality Video Generation" was published, presenting a novel approach to video generation. The VideoFusion model decomposes the diffusion process into two components: base noise and residual noise, which are shared across frames to ensure temporal coherence. By utilizing a pre-trained image diffusion model as a base generator, the model efficiently generated high-quality and coherent videos. Fine-tuning the pre-trained model on video data addressed the domain gap between image and video data, enhancing the model's ability to produce realistic and consistent video sequences. In the same month, Adobe introduced Firefly AI as part of its features. === 2024 === In January 2024, Google announced development of a text-to-video model named Lumiere which is anticipated to integrate advanced video editing capabilities. Matthias Niessner and Lourdes Agapito at AI company Synthesia work on developing 3D neural rendering techniques that can synthesise realistic video by using 2D and 3D neural representations of shape, appearances, and motion for controllable video synthesis of avatars. In June 2024, Luma Labs launched its Dream Machine video tool. That same month, Kuaishou extended its Kling AI text-to-video model to international users. In July 2024, TikTok owner ByteDance released Jimeng AI in China, through its subsidiary, Faceu Technology. By September 2024, the Chinese AI company MiniMax debuted its video-01 model, joining other established AI model companies like Zhipu AI, Baichuan, and Moonshot AI, which contribute to China's involvement in AI technology. In December 2024 Lightricks launched LTX Video as an open source model. === 2025 === Alternative approaches to text-to-video models include Google's Phenaki, Hour One, Colossyan, Runway's Gen-3 Alpha, and OpenAI's Sora, Several additional text-to-video models, such as Plug-and-Play, Text2LIVE, and TuneAVideo, have emerged. FLUX.1 developer Black Forest Labs has announced its text-to-video model SOTA. Google was preparing to launch a video generation tool named Veo for YouTube Shorts in 2025. In May 2025, Google launched the Veo 3 iteration of the model. It was noted for its impressive audio generation capabilities, which were a previous limitation for text-to-video models. In July 2025 Lightricks released an update to LTX Video capable of generating clips reaching 60 seconds, and in October 2025 it released LTX-2, with audio capabilities built in. === 2026 === In February 2026, ByteDance released Seedance 2.0, it was noted for its impressive realistic generation, motion and camera control and 15 second generation, however the model faced huge critiscism from Motion Picture Association for copyright infringement. After viewing a viral clip of a fight between actors Brad Pitt and Tom Cruise, Rhett Reese, who is the co-writer of Deadpool & Wolverine and Zombieland announced that on social media "I hate to say it. It’s likely over for us," further stating that "In next to no time, one person is going to be able to sit at a computer and create a movie indistinguishable from what Hollywood now releases." == Architecture and training == There are several architectures that have been used to create text-to-video models. Similar to text-to-image models, these models can be trained using Recurrent Neural Networks (RNNs) such as long short-term memory (LSTM) networks, which has been used for Pixel Transformation Models and Stochastic Video Generation Models, which aid in consistency and realism respectively. An alternative for these include transformer models. Generative adversarial networks (GANs), Variational autoencoders (VAEs), — which can aid in the prediction of human motion — and diffusion models have also been used to develop the image generation aspects of the model. Text-video datasets used to train models include, but are not limited to, WebVid-10M, HDVILA-100M, CCV, ActivityNet, and Panda-70M. These datasets contain millions of original videos of interest, generated videos, captioned-videos, and textual information that help train models for accuracy. Text-video datasets used to train models include, but are not limited to PromptSource, DiffusionDB, and VidProM. These datasets provide the range of text inputs needed to teach models how to interpret a variety of textual prompts. The video generation process involves synchronizing the text inputs with video frames, ensuring alignment and consistency throughout the sequence. This predictive process is subject to decline in quality as the length of the video increases due to resource limitations. The Will Smith Eating Spaghetti test is a benchmark for models. == Limitations == Despite the rapid evolution of text-to-video models in their performance, a primary limitation is that they are very computationally heavy which limits its capacity to provide high quality and lengthy outputs. Additionally, these models require a large amount of specific training data to be able to generate high quality and coherent outputs, which brings about the issue of accessibility. Moreover, models may misinterpret textual prompts, resulting in video outputs that deviate from the intended meaning. This can occur due to limitations in capturing semantic context embedded in text, which affects the model's ability to align generated video with the user's intended message. Various models, including Make-A-Video, Imagen Video, Phenaki, CogVideo, GODIVA, and NUWA, are currently being tested and refined to enhance their alignment capabilities and overall performance in text-to-video generation. Another issue with the outputs is that text or fine details in AI-generated videos often appear garbled, a problem that stable diffusion models also struggle with. Examples include distorted hands and unreadable text. == Ethics == The deployment of text-to-video models raises ethical considerations related to content generation. These models have the potential to create inappropriate or unauthorized content, including explicit material, graphic violence, misinformation, and likenesses of real individuals without consent. Ensuring that AI-generated content complies with established standards for safe and ethical usage is essential, as content generated by these models may not always be easily identified as harmful or misleading. The ability of AI to recognize and filter out NSFW or copyrighted content remains an ongoing challenge, with implications for both creators and audiences. == Impacts and applications == Text-to-video models offer a broad range of applications that may benefit various fields, from educational and promotional to creative industries. These models can streamline content creation for training videos, movie previews, gaming assets, and visualizations, making it easier to generate content. During the Russo-Ukrainian war, fake videos made with artificial intelligence were created as part of a propaganda war against Ukraine and shared in social media. These included depictions of children in the Ukrainian Armed Forces, fake ads targeting children encouraging them to denounce critics of the Ukrainian government, or fictitious statements by Ukrainian President Volodymyr Zelenskyy about the country's surrender, among others. === Movies === Kaur vs Kore is the first Indian feature film made using generative AI which features dual role for the AI character of Sunny Leone, set to release in 2026. Chiranjeevi Hanuman – The Eternal is an Indian movie made entirely using Generative AI created by Vijay Subramaniam which is set for theatrical release in 2026. The movie was widely criticised by the Film makers in the Bollywood industr
TurboQuant
TurboQuant is an online vector quantization algorithm for compressing high-dimensional Euclidean vectors while preserving their geometric structure. It was proposed in 2025 by Amir Zandieh, Majid Daliri, Majid Hadian, and Vahab Mirrokni in the paper TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate. The paper lists Zandieh and Mirrokni as affiliated with Google Research, Daliri with New York University, and Hadian with Google DeepMind. The method was developed for applications including large language model (LLM) inference, key–value (KV) cache compression, vector databases, and nearest neighbor search. TurboQuant consists of two related algorithms: TurboQuantmse, which is optimized for mean squared error (MSE), and TurboQuantprod, which is optimized for unbiased inner product estimation. The algorithm uses a random rotation of input vectors, applies scalar quantizers to the rotated coordinates, and, for inner-product estimation, applies a one-bit Quantized Johnson–Lindenstrauss (QJL) transform to the residual error. == Background == Vector quantization is a compression method that maps high-dimensional vectors to a finite set of codewords. The problem has roots in Shannon's source coding theory and rate–distortion theory. In machine learning and information retrieval, vector quantization is used to reduce the memory required to store embeddings, activation vectors, and other numerical representations. In Transformer-based large language models, the KV cache stores key and value vectors from previous tokens during autoregressive decoding. The size of this cache grows with context length, the number of attention heads, and the number of concurrent requests, making it a major memory bottleneck in LLM serving. Similar compression problems appear in vector search, where large collections of embedding vectors must be stored and searched efficiently. Earlier approaches to vector quantization include product quantization, scalar quantization, and data-dependent k-means codebook construction. The TurboQuant paper argues that many existing methods either require offline preprocessing and calibration or suffer from suboptimal distortion guarantees in online settings. == Algorithm == === TurboQuantmse === TurboQuantmse is the version of the algorithm optimized for mean-squared error. For a unit vector x ∈ S d − 1 {\displaystyle x\in S^{d-1}} , the algorithm first applies a random rotation matrix Π ∈ R d × d {\displaystyle \Pi \in \mathbb {R} ^{d\times d}} and sets z = Π x {\displaystyle z=\Pi x} . Each coordinate of the rotated vector follows a shifted and scaled beta distribution, which converges to a normal distribution in high dimensions. In high dimensions, distinct coordinates also become nearly independent, allowing the algorithm to apply scalar quantizers independently to each coordinate. The scalar quantizer is constructed by solving a one-dimensional continuous k-means or Lloyd–Max quantization problem. If the centroids are c 1 , c 2 , … , c 2 b {\displaystyle c_{1},c_{2},\ldots ,c_{2^{b}}} , the quantization step stores, for each coordinate, i d x j = a r g m i n k ∈ [ 2 b ] | z j − c k | . {\displaystyle \mathrm {idx} _{j}=\operatorname {} {arg\,min}_{k\in [2^{b}]}|z_{j}-c_{k}|.} During dequantization, the stored index for each coordinate is replaced by the corresponding centroid, giving a reconstructed rotated vector z ~ {\displaystyle {\tilde {z}}} . The algorithm then rotates back: x ~ = Π ⊤ z ~ . {\displaystyle {\tilde {x}}=\Pi ^{\top }{\tilde {z}}.} The paper gives the following bound for TurboQuantmse: D m s e ≤ 3 π 2 ⋅ 1 4 b . {\displaystyle D_{\mathrm {mse} }\leq {\frac {\sqrt {3\pi }}{2}}\cdot {\frac {1}{4^{b}}}.} It also reports finer-grained MSE values of approximately 0.36, 0.117, 0.03, and 0.009 for bit-widths b = 1 , 2 , 3 , 4 {\displaystyle b=1,2,3,4} , respectively. === TurboQuantprod === TurboQuantprod is optimized for unbiased inner-product estimation. The authors note that an MSE-optimized quantizer may introduce bias when used to estimate inner products. To address this, TurboQuantprod first applies TurboQuantmse with bit-width b − 1 {\displaystyle b-1} , then applies a one-bit Quantized Johnson–Lindenstrauss transform to the remaining residual vector. Let r = x − Q m s e − 1 ( Q m s e ( x ) ) {\displaystyle r=x-Q_{\mathrm {mse} }^{-1}(Q_{\mathrm {mse} }(x))} be the residual after MSE quantization, and let γ = ‖ r ‖ 2 {\displaystyle \gamma =\|r\|_{2}} . The QJL step stores a sign vector for the residual. For γ ≠ 0 {\displaystyle \gamma \neq 0} , this can be written using the normalized residual u = r / γ {\displaystyle u=r/\gamma } : q j l = sign ( S u ) , {\displaystyle qjl=\operatorname {sign} (Su),} where S ∈ R d × d {\displaystyle S\in \mathbb {R} ^{d\times d}} is a random projection matrix. Since the sign function is invariant under positive rescaling, this is equivalent to sign ( S r ) {\displaystyle \operatorname {sign} (Sr)} when r ≠ 0 {\displaystyle r\neq 0} . If γ = 0 {\displaystyle \gamma =0} , the residual correction is zero. TurboQuantprod stores the MSE quantization, the QJL sign vector, and the residual norm: Q p r o d ( x ) = [ Q m s e ( x ) , q j l , γ ] . {\displaystyle Q_{\mathrm {prod} }(x)=\left[Q_{\mathrm {mse} }(x),qjl,\gamma \right].} The dequantized vector is reconstructed as x ~ = x ~ m s e + π / 2 d γ S ⊤ q j l . {\displaystyle {\tilde {x}}={\tilde {x}}_{\mathrm {mse} }+{\frac {\sqrt {\pi /2}}{d}}\,\gamma S^{\top }qjl.} The paper proves that TurboQuantprod is unbiased for inner-product estimation: E x ~ [ ⟨ y , x ~ ⟩ ] = ⟨ y , x ⟩ . {\displaystyle \mathbb {E} _{\tilde {x}}\left[\langle y,{\tilde {x}}\rangle \right]=\langle y,x\rangle .} It also gives the distortion bound D p r o d ≤ 3 π 2 ⋅ ‖ y ‖ 2 2 d ⋅ 1 4 b . {\displaystyle D_{\mathrm {prod} }\leq {\frac {\sqrt {3\pi }}{2}}\cdot {\frac {\|y\|_{2}^{2}}{d}}\cdot {\frac {1}{4^{b}}}.} == Performance and applications == The TurboQuant paper reports that the algorithm achieves near-optimal distortion rates within a small constant factor of information-theoretic lower bounds. The authors report that, for KV cache quantization, TurboQuant achieved quality neutrality at 3.5 bits per channel and marginal degradation at 2.5 bits per channel. In long-context LLM experiments using Llama 3.1 8B Instruct, the paper evaluated the method on a "needle-in-a-haystack" retrieval task with document lengths from 4,000 to 104,000 tokens. It reported that TurboQuant matched the uncompressed full-precision baseline while using more than 4× compression, and compared the method against PolarQuant, SnapKV, PyramidKV, and KIVI. Google Research stated that TurboQuant was evaluated on long-context benchmarks including LongBench, Needle in a Haystack, ZeroSCROLLS, RULER, and L-Eval using open-source models including Gemma and Mistral. According to a report in Tom's Hardware, Google described the method as reducing KV-cache memory by at least six times and achieving up to an eightfold improvement in attention-logit computation on Nvidia H100 GPUs compared with unquantized 32-bit keys. TurboQuant has also been applied to nearest-neighbor vector search. The original paper reports experiments on DBpedia entity embeddings and GloVe embeddings, comparing TurboQuant with product quantization and other vector-search quantization baselines. == Relationship to other methods == TurboQuant is related to several methods for efficient large language model inference and high-dimensional search: Product quantization – a vector quantization technique widely used for approximate nearest-neighbor search Quantization (machine learning) – reducing the numerical precision of weights, activations, or cached tensors in machine learning models PagedAttention – a memory-management algorithm for LLM serving that reduces fragmentation in the KV cache Johnson–Lindenstrauss lemma – a result in high-dimensional geometry used in random projection methods Lloyd's algorithm – an algorithm for scalar and vector quantization, including k-means-style codebook construction Unlike PagedAttention, which focuses on memory allocation and cache layout, TurboQuant reduces the numerical storage cost of the vectors themselves. Unlike many product-quantization methods, TurboQuant is designed to be data-oblivious and online, avoiding dataset-specific codebook training. == Limitations == The strongest performance claims for TurboQuant come from the original paper and Google Research's own publication. Coverage in technology media has noted that the broader impact of the method will depend on real-world implementation details, workloads, and hardware architectures.
Apple Intelligence
Apple Intelligence is a generative artificial intelligence system developed by Apple Inc. Relying on a combination of on-device and server processing, it was announced on June 10, 2024, at the 2024 Worldwide Developers Conference, as a built-in feature of Apple's iOS 18, iPadOS 18, and macOS Sequoia, which were announced alongside Apple Intelligence. Apple Intelligence is free for all users with supported devices. On macOS, Apple Intelligence is available only on Apple silicon Mac computers; Intel-based Mac computers are not supported. Features include writing tools that assist users with grammar and proofreading, image generation, summaries of system notifications, AI-assisted image retouching in the Photos app, and integration with ChatGPT, the popular chatbot by OpenAI. As of March 2026, Apple Intelligence is not available yet on devices purchased in mainland China or on any device using an Apple Account set to mainland China, even if the device was bought elsewhere. == History == === Background === Apple first implemented artificial intelligence features in its products with the release of Siri in the iPhone 4S in 2011. In the years after its release, Apple engaged in efforts to ensure its artificial intelligence operations remained covert; according to University of California, Berkeley professor Trevor Darrell, the company's secrecy deterred graduate students. The company started expanding its artificial intelligence team in 2015, opening up its operations by publishing more scientific papers and joining AI industry research groups. Apple reportedly acquired more AI companies from 2016 to 2020. In 2017, Apple released the iPhone 8 and the iPhone X with the A11 Bionic processor, which featured its first dedicated Neural Engine for accelerating common machine learning tasks. Despite its investments in artificial intelligence, Siri was criticized both by reviewers and internally at Apple for lagging behind other AI assistants. The rapid development of generative artificial intelligence and the release of ChatGPT in late 2022 reportedly blindsided Apple executives and forced the company to refocus its efforts on AI. In an interview with Good Morning America, Apple CEO Tim Cook stated that generative AI had "great promise" but had some potential dangers, and that it was "looking closely" at ChatGPT. It was first reported in July 2023 that Apple was creating its own internal large language model, codenamed "Ajax". In October 2023, Apple was reportedly on track to release new generative AI features into its operating systems by 2024, including a significantly redeveloped Siri. In an earnings call in February 2024, Cook stated that the company was spending a "tremendous amount of time and effort" into AI features that would be shared "later that year". === Google deal === In January 2026, Apple and Google announced a multi-year partnership under which Apple’s next-generation foundation models are expected to incorporate Google’s Gemini models and cloud infrastructure. According to the companies, the collaboration is intended to support future Apple Intelligence features, including enhancements to Siri, while Apple Intelligence will continue to operate on Apple devices and through Apple’s Private Cloud Compute system, which Apple states is designed to preserve user privacy. On an earnings call, Apple reported to investors that they were integrating an on-device model of the Google Gemini AI to Siri, as the development of their model was beset with setbacks. Apple has previously tested and used other third-party AI models like ChatGPT, but according to a Bloomberg article by Mark Gurman, Apple pushed forward the proposed Google deal; by using Google's Gemini model possessing 1.2 trillion parameters, Apple would integrate a much larger and more complex model than those it previously developed and used. Of note, comparable AI models from other major companies (including OpenAI and Meta) have also been reported to operate at a similar “trillion-parameter” scale and to compete against Gemini-class systems on benchmarks. == Models == Apple Intelligence consists of an on-device model as well as a cloud model running on servers primarily using Apple silicon. Both models consist of a generic foundation model, as well as multiple adapter models that are more specialized to particular tasks like text summarization and tone adjustment. It was launched for developers and testers on July 29, 2024, in U.S. English, with the developer betas of iOS 18.1, macOS 15.1, and iPadOS 18.1, released partially on October 28, 2024, and will fully launch by 2026. According to a human evaluation done by Apple's machine learning division, the on-device foundation model beat or tied equivalent small models by Mistral AI, Microsoft, and Google, while the server foundation models beat the performance of OpenAI's GPT-3, while roughly matching the performance of GPT-4. Apple's cloud models are built on a Private Cloud Compute platform which is allegedly designed with user privacy and end-to-end encryption in mind. Unlike other generative AI services like ChatGPT which use servers from third parties, Apple Intelligence's cloud models are run entirely on Apple servers with custom Apple silicon hardware built for end-to-end encryption. It was also designed to make sure that the software running on said servers matches the independently verifiable software accessible to researchers. In case of a software mismatch, Apple devices will refuse to connect to the servers. On June 10, 2025, Apple announced that Apple's on-device foundation models will be available to third-party applications as part of the Foundation Models API, with support for structured data response and tool calling. == Features == === Writing tools === Apple Intelligence features writing tools that are powered by LLMs. Selected text can be proofread, rewritten, made more friendly, concise or professional, similar to the AI writing features of the popular online English-language writing assistant tool Grammarly. It can also be used to generate summaries, key points, tables, and lists from an article or piece of writing. In iOS 18.2 and macOS 15.2, a ChatGPT integration was added to Writing Tools through "Compose" and "Describe your change" features. === Real-time Translation === Apple Intelligence enables the real-time translation of messages, photos and videos, and phone calls, through Apple's hardware. For communicating with foreigners, using the Translate app on iPhone to show subtitles in their language or to play back the translated audio naturally in their language, and also by wearing AirPods with Live Translation can now help to understand what someone is saying in users' preferred language in conversation. If both have headphones, simultaneous interpretation can be achieved. === Image Playground === Apple Intelligence can be used to generate images on-device with the Image Playground app. Similarly to OpenAI's DALL-E, it can be used to generate images using AI, using phrases and descriptions to output an image with customizable styles such as Animation and Sketch. In Notes, users can access Image Playground on iPad through the Image Wand tool in the Apple Pencil palette without having to open the Image Playground app. Rough sketches made with Apple Pencil can be transformed into images. As part of iOS, iPadOS, and macOS 26, Image Playground now integrates with the image generation models built into ChatGPT. === Genmoji === Using Apple Intelligence text-to-image models, users can generate unique "Genmoji" images by typing descriptions (prompting). Users can pick people in photos to have Genmoji generate images that resemble them. Similarly to emoji, Genmoji can be added inline to text messages, tapbacks, stickers and can be shared in Messages as well in third-party applications as inline messages or as stickers. === Siri overhaul === Siri, which used to be Apple's virtual assistant, has been updated to be an LLM chatbot, with enhanced capabilities made possible by Apple Intelligence. The latest iteration features an updated user interface, improved natural language processing, and the option to interact via text by double tapping the home bar without enabling the feature in the Accessibility menu, or double-clicking the command key on macOS. In a later update, Apple Intelligence will add the ability for Siri to use personal context from device activities to answer queries. === Mail === Apple Intelligence adds a feature called Priority Messages to the Mail app, which shows urgent emails such as same-day invitations or boarding passes, with AI generated summaries of the email. The Mail app also gains the ability to categorize incoming mail into Primary, Transactions, Updates, and Promotions based on what the email contains, which Apple claims is done all on-device. === Photos === Apple's Photos app includes a feature to create custom memory movies and enhanced search capabilities. Users can describe
Pol.is
Polis (or Pol.is) is wiki survey software designed for large group collaborations. As a civic technology, Polis allows people to share their opinions and ideas, and its algorithm is intended to elevate ideas that can facilitate better decision-making, especially when there are lots of participants. Polis has been credited for assisting the passage of legislation in Taiwan. Pol.is has been used by governments in the United States, Canada, Singapore, Philippines, Finland, Spain and elsewhere. == History == Pol.is was founded by Colin Megill, Christopher Small, and Michael Bjorkegren after the Occupy Wall Street and Arab Spring movements. In Taiwan, pol.is has been "one of the key parts" of vTaiwan's suite of open-source tools for its citizen engagement efforts arising out of the Sunflower Student Movement. vTaiwan claims that of the 26 national issues related to technology discussed on the platform, 80% led to government action. Pol.is is also utilized by "Join," a national platform for online deliberation run by the Taiwanese government. In 2022, Wired reported that Polis was an influence on the Community Notes project at Twitter. In 2023, Megill advised OpenAI on how to facilitate deliberation at scale in a way that was more efficient than Polis, which still required significant human labor and analysis at the time. He helped to award $1 million in grants to teams working on solving the problem of deliberation at scale. In 2023, Anthropic was also exploring steering model behavior using Polis. In 2025, it helped the county that includes Bowling Green, Kentucky make a 25 year plan by facilitating the collection and review of ideas from thousands of residents, representing 10% of the county. 2,370 of 3,940 unique ideas were agreed-upon by over 80% of survey respondents. Ideas were screened by volunteers if they were redundant to an existing idea, off-topic or obscene. == How it works == Pol.is participants are anonymous and cannot reply directly to others posts, in an effort to avoid personal attacks for users. Its algorithms are designed not for engagement and scrolling, but to find areas of agreement to better understand the nuances of a wide range of opinions. Participants are prompted for ideas and vote on other participants' ideas. == Reception == Andrew Leonard, The Financial Times, and VentureBeat describe Pol.is as a possible antidote to the divisiveness of traditional internet discourse by gamifying consensus. Audrey Tang agreed saying, "Polis is quite well known in that it's a kind of social media that instead of polarizing people to drive so called engagement or addiction or attention, it automatically drives bridge making narratives and statements. So only the ideas that speak to both sides or to multiple sides will gain prominence in Polis." Niall Ferguson argues that the approach to utilize tools like Pol.is and Join in Taiwan empowers ordinary people instead of the elite and protects individual freedoms, providing a contrast to the AI-enhanced panopticon model seen in China. Carl Miller praised the technology as having "gamified finding consensus." Darshana Narayanan, in an op-ed in the Economist, argues that open-source machine-learning-based tools like Polis can help to bypass the influence of special interests or experts. Jamie Susskind cited polis and vTaiwan as a model for democracies, particularly around digital policy issues.
User-defined function
A user-defined function (UDF) is a function provided by the user of a program or environment, in a context where the usual assumption is that functions are built into the program or environment. UDFs are usually written for the requirement of its creator. == BASIC language == In some old implementations of the BASIC programming language, user-defined functions are defined using the "DEF FN" syntax. More modern dialects of BASIC are influenced by the structured programming paradigm, where most or all of the code is written as user-defined functions or procedures, and the concept becomes practically redundant. == COBOL language == In the COBOL programming language, a user-defined function is an entity that is defined by the user by specifying a FUNCTION-ID paragraph. A user-defined function must return a value by specifying the RETURNING phrase of the procedure division header and they are invoked using the function-identifier syntax. See the ISO/IEC 1989:2014 Programming Language COBOL standard for details. As of May 2022, the IBM Enterprise COBOL for z/OS 6.4 (IBM COBOL) compiler contains support for user-defined functions. == Databases == In relational database management systems, a user-defined function provides a mechanism for extending the functionality of the database server by adding a function, that can be evaluated in standard query language (usually SQL) statements. The SQL standard distinguishes between scalar and table functions. A scalar function returns only a single value (or NULL), whereas a table function returns a (relational) table comprising zero or more rows, each row with one or more columns. User-defined functions in SQL are declared using the CREATE FUNCTION statement. For example, a user-defined function that converts Celsius to Fahrenheit (a temperature scale used in USA) might be declared like this: Once created, a user-defined function may be used in expressions in SQL statements. For example, it can be invoked where most other intrinsic functions are allowed. This also includes SELECT statements, where the function can be used against data stored in tables in the database. Conceptually, the function is evaluated once per row in such usage. For example, assume a table named Elements, with a row for each known chemical element. The table has a column named BoilingPoint for the boiling point of that element, in Celsius. The query would retrieve the name and the boiling point from each row. It invokes the CtoF user-defined function as declared above in order to convert the value in the column to a value in Fahrenheit. Each user-defined function carries certain properties or characteristics. The SQL standard defines the following properties: Language - defines the programming language in which the user-defined function is implemented; examples include SQL, C, C# and Java. Parameter style - defines the conventions that are used to pass the function parameters and results between the implementation of the function and the database system (only applicable if language is not SQL). Specific name - a name for the function that is unique within the database. Note that the function name does not have to be unique, considering overloaded functions. Some SQL implementations require that function names are unique within a database, and overloaded functions are not allowed. Determinism - specifies whether the function is deterministic or not. The determinism characteristic has an influence on the query optimizer when compiling a SQL statement. SQL-data access - tells the database management system whether the function contains no SQL statements (NO SQL), contains SQL statements but does not access any tables or views (CONTAINS SQL), reads data from tables or views (READS SQL DATA), or actually modifies data in the database (MODIFIES SQL DATA). User-defined functions should not be confused with stored procedures. Stored procedures allow the user to group a set of SQL commands. A procedure can accept parameters and execute its SQL statements depending on those parameters. A procedure is not an expression and, thus, cannot be used like user-defined functions. Some database management systems allow the creation of user defined functions in languages other than SQL. Microsoft SQL Server, for example, allows the user to use .NET languages including C# for this purpose. DB2 and Oracle support user-defined functions written in C or Java programming languages. === SQL Server 2000 === There are three types of UDF in Microsoft SQL Server 2000: scalar functions, inline table-valued functions, and multistatement table-valued functions. Scalar functions return a single data value (not a table) with RETURNS clause. Scalar functions can use all scalar data types, with exception of timestamp and user-defined data types. Inline table-valued functions return the result set of a single SELECT statement. Multistatement table-valued functions return a table, which was built with many TRANSACT-SQL statements. User-defined functions can be invoked from a query like built‑in functions such as OBJECT_ID, LEN, DATEDIFF, or can be executed through an EXECUTE statement like stored procedures. Performance Notes: User-defined functions are subroutines made of one or more Transact-SQL statements that can be used to encapsulate code for reuse. It takes zero or more arguments and evaluates a return value. Has both control-flow and DML statements in its body similar to stored procedures. Does not allow changes to any Global Session State, like modifications to database or external resource, such as a file or network. Does not support output parameter. DEFAULT keyword must be specified to pass the default value of parameter. Errors in UDF cause UDF to abort which, in turn, aborts the statement that invoked the UDF. === Apache Hive === Apache Hive defines, in addition to the regular user-defined functions (UDF), also user-defined aggregate functions (UDAF) and table-generating functions (UDTF). Hive enables developers to create their own custom functions with Java. === Apache Doris === Apache Doris, an open-source real-time analytical database, allows external users to contribute their own UDFs written in C++ to it.
Harold Borko
Harold Borko (1922-2012) was an American psychologist and researcher working primarily in the field of information science. == Biography == Borko was born in 1922 in New York City, New York. After serving in the US Army from 1942 to 1946 he obtained a BA in Psychology from the University of California, Los Angeles in 1948 and both his MA and PhD from the University of Southern California in Psychology in 1952. He returned to the army as a psychologist until 1956 after which he began a career working in and teaching information science. He died in California in 2012. == Information Science Career == After leaving the military Borko began working at the RAND Corporation as a Systems Training Specialist in 1956 and moved to the Systems Development Corporation a year later working in the Language Processing and Retrieval department. Alongside this work he taught Psychology at USC from 1957-65 and then moved into teaching Library Science at UCLA from 1965. In 1967 Borko left his role at the Systems Development Corporation and continued as a full-time professor at UCLA until his retirement in 1993.. From 1961 to 1995 Borko authored and co-authored over 100 articles on new developments in the field as well as the historiography of information science. He served as an editor of the Journal of Educational Data Processing from 1963-1975 and as President of the American Society for Information Science in 1966 == Partial list of works == Borko, H. (1962, May). The construction of an empirically based mathematically derived classification system. In Proceedings of the May 1-3, 1962, spring joint computer conference (pp. 279-289). Borko, H., & Bernick, M. (1963). Automatic document classification. Journal of the ACM (JACM), 10(2), 151-162. Borko, H. (1964). The Storage and Retrieval of Educational Information. Journal of Teacher Education, 15(4), 449-452. Borko, H. (1964). Measuring the reliability of subject classification by men and machines. American Documentation, 15(4), 268-273. Borko, H. (1965). The conceptual foundations of information systems. Borko, H. (1968), Information science: What is it?†. Amer. Doc., 19: 3-5. https://doi.org/10.1002/asi.5090190103 Borko, H. (1970). Experiments in book indexing by computer. Information storage and retrieval, 6(1), 5-16. Borko, H. (1985). An introduction to computer-based library systems (Lucy A. Tedd). Education for Information, 3(1), 61.