AI’s promise to indie filmmakers: Faster, cheaper, lonelier

12:21 AM IST · February 21, 2026

A Filipino man walks through the backyard of his childhood home in rural Hawai’i, his footsteps swooshing through the grass. Birds chirp, contributing to the tropical din, as he approaches a shrine at the base of a starfruit tree. He bends to inspect a framed black-and-white photograph of a woman, her hair in a 1950s side part. Suddenly, a gust of wind shakes the tree’s branches, knocking over the contents of the shrine. The man steps back, trips on a root, and hits his head. When he awakens, he’s in a dark, misty forest, a woman wearing a clay mask standing over him, brandishing a sword. “Who are you who dares to sleep under the sacred tree?” she asks in Ilocano, a Philippine language widely spoken in Hawaii’s Filipino community, while holding the sword at his throat. He replies that he’s lost and turns to flee. She chases, alternating between running and floating through the air. He falls again. She advances, sword held high. He throws a rock at her, shattering the clay mask and revealing half her face. “Mom?” he asks. This is the opening of “Murmuray,” a short film by independent filmmaker Brad Tangonan. Everything about this film felt like his previous work, from the tactile nature shots to the dreamlike desaturated highlights. The only difference? He made it using AI. Tangonan was one of 10 filmmakers to participate in Google Flow Sessions, a five-week cohort that gave creatives access to Google’s suite of AI tools to produce short films, including Gemini, image generator Nano Banana Pro, and film generator Veo. Each film differed in scope. Hal Watmough’s “You’ve Been Here Before” blended hyperreal, lifelike visuals with cartoonish stylization to playfully explore the importance of a morning routine, while Tabitha Swanson’s“The Antidote to Fear is Curiosity”is a more esoteric, philosophical conversation about our relationship with AI and ourselves. None of these short films, which were screened at Soho House New York late last year, felt like AI slop. Each independent filmmaker I spoke to said that, in the case of these films, AI had enabled them to tell a story they otherwise wouldn’t have had the budget or time to tell. “I see all of these tools, whether it be a camera you can pick up or generative AI, as ways for an artist to express what they have in their mind,” Tangonan told me after the screenings. This AI-is-just-another-tool-for-creators argument is certainly the message Google is trying to underscore. Google isn’t wrong; AI will increasingly be part of a creator’s toolkit as video generation products improve. In 2025, companies like Google, Runway, OpenAI, Kling, Luma AI, and Higgsfield progressed far beyond the uncanny, prompt-based novelties of the year prior. The AI video industry, with billions in venture capital dollars in tow, is now moving from prototype to post-production. This era of AI abundance that has provided tools to “democratize access” to the film industry also threatens to erase jobs and creativity, smothering them under an avalanche of low-effort slop. The existential stakes have pitted creatives against one another. Those who engage with AI risk being labeled as complicit; those who don’t risk becoming obsolete. The question isn’t whether the tools belong in the toolkit — they’re coming, whether we like it or not. Instead it is: What kind of filmmaking survives when the industry pushes for speed and scale over quality? And what happens when individual artists use the same tools to make something that actually matters? The arguments against AI in filmmaking are plentiful — and from some of the highest-profile names in the industry. Filmmaker Guillermo del Torosaid last Octoberthat he would rather die than use generative AI to make a film. James Cameron said in a recentCBS interviewthe idea of generating actors and emotions with prompts is “horrifying,” and that generative AI is only capable of spitting out a blended average of everything that’s ever been done by humans before. Werner Herzog said the films he’s seen created by AI “have no soul.” He added: “The common denominator, and nothing beyond this common denominator, can be found in these fabrications.” Cameron and Herzog’s thesis is that AI is taking the wheel of creation out of the hands of humans and couldn’t possibly be used to create a representation of their own lived experiences. “It’s very easy to be angry with AI as a concept in the machine, but it’s harder to be angry with someone that’s made something personal,” Watmough told TechCrunch. Tangonan, who describes “Murmuray” as a “family story,” agrees with that sentiment. “AI is a facilitator,” Tangonan said. “I’m still making all the creative decisions. When people see ‘AI slop’ online, it’s a lot of lowest common denominator stuff. And, yeah, if you hand over the keys to AI, that’s what you’re going to get. But if you have a voice and a creative perspective and a style, then you’re going to get something different.” Using AI in filmmaking doesn’t mean just prompting a film into existence. Tangonan, for example, wrote the script for “Murmuray” without AI and gathered visual references for a shot list. He then fed that content into Nano Banana Pro to generate images that matched his style and served as the foundation for video generation. Filmmaker Keenan MacWilliam also took pains to ensure her short film “Mimesis,” a fictional guided meditation, was a “true extension of [her] visual language, rather than a ‘blender’ of other artists’ work.” MacWilliam wrote the script and recorded her own voice for the mock meditation, which was equal parts relaxing and funny. Onscreen, over a black, watery backdrop, psychedelic images of flowers and plants blended into each other, turned into smoke, morphed into seahorses, and swam away. The images all came from MacWilliam’s own collection of scanned flora and fauna — she travels with her scanner everywhere she goes. “I spent a lot of time learning how to make apps that were built with my own dataset, and then used those as reference points,” MacWilliam told TechCrunch, adding that she worked with her long-time composer and sound designer on the film. “I made a choice to avoid using AI for anything that I could have shot with a camera or ask my collaborators to animate. My goal was to unlock new forms of expression for my established themes and style, not to replace the roles of the people who I like to work with.” That was a common thread among the filmmakers I spoke to at the Google Flow event — the desire to use AI only in cases when it was not possible to rely on other humans, or when the strange nature of AI generations serve the story. For example, Sander van Bellegem’s “Melongray” explored the acceleration of life through trippy visualizations. In one shot, a salamander transforms into a balloon. It wasn’t part of his original storyline, but he was inspired by the way AI allowed him to push the limits of both his imagination and physics. Today’s film studiobudgets are being squeezedby rising filming costs, thepivot to streaming, and risk-averse corporate consolidation. That means big spends are saved for predictable revenue generators (see: the millionth Marvel movie) and originalmid-budget movieshave all but been abandoned. Adding AI to the mix risks exacerbating the scarcity mindset of studios to the point where they might try to replace anything that can be — actors, sets, lighting — art and quality be damned. However, the efficiencies AI brings could also lower barriers and make it easier for film studios to produce original work. Even Cameron noted in his CBS interview that generative AI could make VFX cheaper, which could lead to more imaginative sci-fi and fantasy films — expensive endeavors that are reserved for existing IP like “Avatar.” The shot in “Murmuray” where the woman is flying through the forest would have taken expensive VFX or very complex rigging on set, both out of budget for a short film, according to Tangonan. But even filmmakers who see the benefits in efficiency understand the risks to artistic expression. “I think efficiency in general is not the best friend of creativity,” MacWilliam said. For independent filmmakers, having so many powerful tools at their disposal is a blessing and a curse. It “democratizes access,” sure, but it also means working alone. The more youcando yourself, the less reason there is to collaborate. “I know I’m a one-man band, and I just made all this by myself…but that should never be the way that anyone tells a story or makes a film,” Watmough told TechCrunch, noting that an actor friend of his contributed the voice for his short. “It should be a collaborative process because the more people that are involved, the more accessible it is by everyone and the more it reaches and connects with people.” Directors make creative decisions, but not all of them. The filmmakers I spoke to found themselves suddenly playing set designer, lighting director, costumer — roles requiring expertise they didn’t have. It was frustrating and draining, pulling them away from the work they actually cared about. And upsetting to think about how an entire ecosystem could be upended so swiftly. The filmmakers I spoke to also said they’d rather not replace actors with AI, though some said AI-generated actors are an inevitability for smaller studios. The tools exist, and are increasingly getting better, to generate actors, their emotions, their movements. AI video startups like Luma AI, which last November raiseda $900 million Series C, are even building technology that allows you to shoot an actor’s performance once, and then use AI to change the character, costume, and set. “In an ideal world, I would work with real actors and some cinematographer and department heads and the full crew to make something amazing and use AI and complement that to be able to do things that we can’t do on set, whether for budgetary or time reasons,” Tangonan. “I think making any creative work that uses new technology always requires a certain kind of gut check and a willingness to have conversations around the work,” MacWilliam said. “These are tools,” she added. “How are you going to use the tool? Are you going to be ethical about it? Are you going to ask questions? Are you going to be transparent and share knowledge?” But many don’t see AI tools as neutral. Labor replacement aside, there are still copyright concerns. AI video generation startupRunwayhas reportedly scraped thousands of hours of YouTube videos and copyrighted studio content, while others — including Google, OpenAI, and Luma AI — have faced questions about whether they are doing the same, or training on copyrighted films and stock footage without permission. (Though some tools, likeMoonvalley’s Marey, are trained only on openly licensed data.) Then there are the environmental horrors —some estimates suggestgenerating seconds of AI video can consume as much electricity as hours of streaming. Unsurprisingly, many of the filmmakers I spoke to said they face stigma for experimenting with AI. “Whenever I do post things online, a lot of my filmmaking colleagues have a very knee-jerk reaction to it that we should all hold the line and not use any of these tools,” Tangonan said. “I just don’t agree with that.” If filmmakers are too afraid to discuss how AI can and should be used and what the ethical boundaries are, then the conversation risks being decided for them. Not by artists trying to use it responsibly, but by efficiency-crazed studios that care more about bottom lines than art. “The film industry is floundering because people aren’t innovating and everything costs too much. We need tools like this for it to survive,” said Watmough. “I think it’s essential that people engage with it because if we don’t, then it’s going to become something we don’t recognize, and that’s not sustainable.” Correction: An earlier version of this article mischaracterized Ilocano as a Hawaiian dialect of Filipino. Ilocano is a language from the northern Philippines and is widely spoken among Filipino communities in Hawaii.

Latest AI News

View All News →

In Swiggy's Developer Community, Some are Planning Parties, Other are Prepping Meals

Swiggy is enabling developers to build health assistants, accessibility tools, and conversational shopping agents.

6 hours ago

View

The only AI glossary you’ll need this year

Artificial intelligence is rewriting the world, and simultaneously inventing a whole new language to describe how it’s doing it. Sit in on any product meeting, pitch, or panel these days, and you’ll hear people toss around LLMs, RAG, RLHF, and a dozen other terms that can make even very smart people in the tech world feel a little insecure. This glossary is our attempt to fix that: pain-English definitions of the AI terms you’re most likely to actually run into, whether you’re building with this stuff, investing in it, or just trying to keep up by reading TechCrunch or listening to related podcasts. We update it regularly as the field evolves, so consider it a living document, much like the AI systems it describes. Artificial general intelligence, or AGI, is a nebulous term. But it generally refers to AI that’s more capable than the average human at many, if not most, tasks. OpenAI CEO Sam Altman once described AGI as the “equivalent of a median human that you couldhire as a co-worker.” Meanwhile,OpenAI’s charterdefines AGI as “highly autonomous systems that outperform humans at most economically valuable work.” Google DeepMind’s understanding differs slightly from these two definitions; the lab views AGI as “AI that’s at least as capable as humans at most cognitive tasks.” Confused? Not to worry —so are experts at the forefront of AI research. An AI agent refers to a tool that uses AI technologies to perform a series of tasks on your behalf — beyond what a more basic AI chatbot could do — such as filing expenses, booking tickets or a table at a restaurant, or even writing and maintaining code. However, as we’veexplained before, there are lots of moving pieces in this emergent space, so “AI agent” might mean different things to different people. Infrastructure is also still being built out to deliver on its envisaged capabilities. But the basic concept implies an autonomous system that may draw on multiple AI systems to carry out multistep tasks. Think of API endpoints as “buttons” on the back of a piece of software that other programs can press to make it do things. Developers use these interfaces to build integrations — for example, allowing one application to pull data from another, or enabling an AI agent to control third-party services directly without a human manually operating each interface. Most smart home devices and connected platforms have these hidden buttons available, even if ordinary users never see or interact with them. As AI agents grow more capable, they are increasingly able to find and use these endpoints on their own, opening up powerful — and sometimes unexpected — possibilities for automation. Given a simple question, a human brain can answer without even thinking too much about it — things like “which animal is taller, a giraffe or a cat?” But in many cases, you often need a pen and paper to come up with the right answer because there are intermediary steps. For instance, if a farmer has chickens and cows, and together they have 40 heads and 120 legs, you might need to write down a simple equation to come up with the answer (20 chickens and 20 cows). In an AI context, chain-of-thought reasoning for large language models means breaking down a problem into smaller, intermediate steps to improve the quality of the end result. It usually takes longer to get an answer, but the answer is more likely to be correct, especially in a logic or coding context. Reasoning models are developed from traditional large language models and optimized for chain-of-thought thinking thanks to reinforcement learning. (See:Large language model) This is a more specific concept that an “AI agent,” which means a program that can take actions on its own, step by step, to complete a goal. A coding agent is a specialized version applied to software development. Rather than simply suggesting code for a human to review and paste in, a coding agent can write, test, and debug code autonomously, handling the kind of iterative, trial-and-error work that typically consumes a developer’s day. These agents can operate across entire codebases, spotting bugs, running tests, and pushing fixes with minimal human oversight. Think of it like hiring a very fast intern who never sleeps and never loses focus — though, as with any intern, a human still needs to review the work. Although somewhat of a multivalent term, compute generally refers to the vitalcomputational powerthat allows AI models to operate. This type of processing fuels the AI industry, giving it the ability to train and deploy its powerful models. The term is often a shorthand for the kinds of hardware that provides the computational power — things like GPUs, CPUs, TPUs, and other forms of infrastructure that form the bedrock of the modern AI industry. A subset of self-improving machine learning in which AI algorithms are designed with a multi-layered, artificial neural network (ANN) structure. This allows them to make more complex correlations compared to simpler machine learning-based systems, such as linear models or decision trees. The structure of deep learning algorithms draws inspiration from the interconnected pathways of neurons in the human brain. Deep learning AI models are able to identify important characteristics in data themselves, rather than requiring human engineers to define these features. The structure also supports algorithms that can learn from errors and, through a process of repetition and adjustment, improve their own outputs. However, deep learning systems require a lot of data points to yield good results (millions or more). They also typically take longer to train compared to simpler machine learning algorithms — so development costs tend to be higher. (See:Neural network) Diffusion is the tech at the heart of many art-, music-, and text-generating AI models. Inspired by physics,diffusion systems slowly “destroy” the structure of data— for example, photos, songs, and so on — by adding noise until there’s nothing left. In physics, diffusion is spontaneous and irreversible — sugar diffused in coffee can’t be restored to cube form. But diffusion systems in AI aim to learn a sort of “reverse diffusion” process to restore the destroyed data, gaining the ability to recover the data from noise. Distillation is a technique used to extract knowledge from a large AI model with a ‘teacher-student’ model. Developers send requests to a teacher model and record the outputs. Answers are sometimes compared with a dataset to see how accurate they are. These outputs are then used to train the student model, which is trained to approximate the teacher’s behavior. Distillation can be used to create a smaller, more efficient model based on a larger model with a minimal distillation loss. This is likely how OpenAI developed GPT-4 Turbo, a faster version of GPT-4. While all AI companies use distillation internally, it may have also been used by some AI companies to catch up with frontier models. Distillation from a competitor usuallyviolatesthe terms of service of AI API and chat assistants. This refers to the further training of an AI model to optimize performance for a more specific task or area than was previously a focal point of its training — typically by feeding in new, specialized (i.e., task-oriented) data. Many AI startups are taking large language models as a starting point to build a commercial product but are vying to amp up utility for a target sector or task by supplementing earlier training cycles with fine-tuning based on their own domain-specific knowledge and expertise. (See:Large language model [LLM]) A GAN, or Generative Adversarial Network, is a type of machine learning framework that underpins some important developments in generative AI when it comes to producing realistic data — including (but not only) deepfake tools. GANs involve the use of a pair of neural networks, one of which draws on its training data to generate an output that is passed to the other model to evaluate. The two models are essentially programmed to try to outdo each other. The generator is trying to get its output past the discriminator, while the discriminator is working to spot artificially generated data. This structured contest can optimize AI outputs to be more realistic without the need for additional human intervention. Though GANs work best for narrower applications (such as producing realistic photos or videos), rather than general purpose AI. Hallucination is the AI industry’s preferred term for AI models making stuff up — literally generating information that is incorrect. Obviously, it’s a huge problem for AI quality. Hallucinations produce GenAI outputs that can be misleading and could even lead to real-life risks — with potentially dangerous consequences (think of a health query that returns harmful medical advice). The problem of AIs fabricating information is thought to arise as a consequence of gaps in training data. Hallucinations are contributing to a push toward increasingly specialized and/or vertical AI models — i.e. domain-specific AIs that require narrower expertise — as a way to reduce the likelihood of knowledge gaps and shrink disinformation risks. Inference is the process of running an AI model. It’s setting a model loose to make predictions or draw conclusions from previously seen data. To be clear, inference can’t happen without training; a model must learn patterns in a set of data before it can effectively extrapolate from this training data. Many types of hardware can perform inference, ranging from smartphone processors to beefy GPUs to custom-designed AI accelerators. But not all of them can run models equally well. Very large models would take ages to make predictions on, say, a laptop versus a cloud server with high-end AI chips. [See:Training] Large language models, or LLMs, are the AI models used by popular AI assistants, such asChatGPT,Claude,Google’s Gemini,Meta’s AI Llama,Microsoft Copilot, orMistral’s Le Chat. When you chat with an AI assistant, you interact with a large language model that processes your request directly or with the help of different available tools, such as web browsing or code interpreters. LLMs are deep neural networks made of billions of numerical parameters (or weights, see below) that learn the relationships between words and phrases and create a representation of language, a sort of multidimensional map of words. These models are created from encoding the patterns they find in billions of books, articles, and transcripts. When you prompt an LLM, the model generates the most likely pattern that fits the prompt. (See:Neural network) Memory cache refers to an important process that boosts inference (which is the process by which AI works to generate a response to a user’s query). In essence, caching is an optimization technique, designed to make inference more efficient. AI is obviously driven by high-octane mathematical calculations and every time those calculations are made, they use up more power. Caching is designed to cut down on the number of calculations a model might have to run by saving particular calculations for future user queries and operations. There are different kinds of memory caching, although one of the more well-known isKV (or key value) caching. KV caching works in transformer-based models, and increases efficiency, driving faster results by reducing the amount of time (and algorithmic labor) it takes to generate answers to user questions. (See:Inference) Model Context Protocol, or MCP, is an open standard that lets AI models connect to outside tools and data — your files, databases, or apps like Slack and Google Drive — without a developer building a custom connector for every single pairing. Think of it as a USB-C port for AI. Anthropic introduced MCP in 2024 and later handed it over to the Linux Foundation, and it’s since been adopted by OpenAI, Google, and Microsoft, making it one of the fastest-spreading standards in recent AI history. Mixture of Experts is a model architecture that splits a neural network into many smaller specialized sub-networks, or “experts,” and only activates a handful of them for any given task. Rather than routing every request through the entire model — like calling in your whole office for every question — an MoE model has a built-in “router” that picks just the right specialists for the job. This makes it possible to build enormous models that stay relatively fast and cheap to run, since only a fraction of the network is doing work at any one time. Mistral AI’s Mixtral model is a well-known example; OpenAI’s newer GPT models are also widely believed to use some version of this approach, though the company has never officially confirmed it. (See:Neural network,Deep learning) A neural network refers to the multi-layered algorithmic structure that underpins deep learning — and, more broadly, the whole boom in generative AI tools following the emergence of large language models. Although the idea of taking inspiration from the densely interconnected pathways of the human brain as a design structure for data processing algorithms dates all the way back to the 1940s, it was the much more recent rise of graphical processing hardware (GPUs) — via the video game industry — that really unlocked the power of this theory. These chips proved well suited to training algorithms with many more layers than was possible in earlier epochs — enabling neural network-based AI systems to achieve far better performance across many domains, including voice recognition, autonomous navigation, and drug discovery. (See:Large language model [LLM]) Open source refers to software — or, increasingly, AI models — where the underlying code is made publicly available for anyone to use, inspect, or modify. In the AI world, Meta’s Llama family of models is a prominent example; Linux is the famous historical parallel in operating systems. Open source approaches allow researchers, developers, and companies around the world to build on top of one another’s work, accelerating progress and enabling independent safety audits that closed systems cannot easily provide. Closed source means the code is private — you can use the product but not see how it works, as is the case with OpenAI’s GPT models — a distinction that has become one of the defining debates in the AI industry. Parallelization means doing many things at the same time instead of one after another — like having 10 employees working on different parts of a project at the same time instead of one employee doing everything sequentially. In AI, parallelization is fundamental to both training and inference: modern GPUs are specifically designed to perform thousands of calculations in parallel, which is a big reason why they became the hardware backbone of the industry. As AI systems grow more complex and models grow larger, the ability to parallelize work across many chips and many machines has become one of the most important factors in determining how quickly and cost-effectively models can be built and deployed. Research into better parallelization strategies is now a field of study in its own right. RAMageddon is the fun new term for a not-so-fun trend that is sweeping the tech industry: an ever-increasing shortage of random access memory, or RAM chips, which power pretty much all the tech products we use in our daily lives. As the AI industry has blossomed, the biggest tech companies and AI labs — all vying to have the most powerful and efficient AI — are buying so much RAM to power their data centers that there’s not much left for the rest of us. And that supply bottleneck means that what’s left is getting more and more expensive. That includes industries like gaming (where major companies have had toraise prices on consolesbecause it’s harder to find memory chips for their devices), consumer electronics (where memory shortage could causethe biggest dip in smartphone shipmentsin more than a decade), and general enterprise computing (because those companies can’t get enough RAM for their own data centers). The surge in prices is only expected to stop after the dreaded shortage ends but, unfortunately, there’snot really much of a signthat’s going to happen anytime soon. Like AGI, recursive self-improvement is a threshhold for how smart AI can get, and how little it may rely on humans. In the RSI scenario, AI models start improving themselves without human intervention, leading to a huge acceleration in capabilities and autonomy. In some tellings, this would be a cataclysmic moment akin to the singularity, a moment when AI models become immune to outside intervention. But RSI also describes a basic capability — can an AI model design its own successor? — which makes it much easier for engineers to try to build it.A number of recent AI startupshave set out to build recursively self-improving models, but most of them dismiss the apocalyptic implications, presenting RSI as simply the next frontier for research. Reinforcement learning is a way of training AI where a system learns by trying things and receiving rewards for correct answers — like training your beloved pet with treats, except the “pet” in this scenario is a neural network and the “treat” is a mathematical signal indicating success. Unlike supervised learning, where a model is trained on a fixed dataset of labeled examples, reinforcement learning lets a model explore its environment, take actions, and continuously update its behavior based on the feedback it receives. This approach has proven especially powerful for training AI to play games, control robots, and, more recently, sharpen the reasoning ability of large language models. Techniques like reinforcement learning from human feedback, or RLHF, are now central to how leading AI labs fine-tune their models to be more helpful, accurate, and safe. When it comes to human-machine communication, there are some obvious challenges — people communicate using human language, while AI programs execute tasks through complex algorithmic processes informed by data. Tokens bridge that gap: they are the basic building blocks of human-AI communication, representing discrete segments of data that have been processed or produced by an LLM. They are created through a process called tokenization, which breaks down raw text into bite-sized units a language model can digest, similar to how a compiler translates human language into binary code a computer can understand. In enterprise settings, tokens also determine cost — most AI companies charge for LLM usage on a per-token basis, meaning the more a business uses, the more it pays. So again, tokens are the small chunks of text — often parts of words rather than whole ones — that AI language models break language into before processing it; they are roughly analogous to “words” for the purposes of understanding AI workloads. Throughput refers to how much can be processed in a given period of time, so token throughput is essentially a measure of how much AI work a system can handle at once. High token throughput is a key goal for AI infrastructure teams, since it determines how many users a model can serve simultaneously and how quickly each of them receives a response. AI researcher Andrej Karpathy has described feeling anxious when his AI subscriptions sit idle — echoing the feeling he had as a grad student when expensive computer hardware wasn’t being fully utilized — a sentiment that captures why maximizing token throughput has become something of an obsession in the field. Developing machine learning AIs involves a process known as training. In simple terms, this refers to data being fed in in order that the model can learn from patterns and generate useful outputs. Essentially, it’s the process of the system responding to characteristics in the data that enables it to adapt outputs toward a sought-for goal — whether that’s identifying images of cats or producing a haiku on demand. Training can be expensive because it requireslotsof inputs, and the volumes required have been trending upwards — which is why hybrid approaches, such as fine-tuning a rules-based AI with targeted data, can help manage costs without starting entirely from scratch. [See:Inference] A technique where a previously trained AI model is used as the starting point for developing a new model for a different but typically related task — allowing knowledge gained in previous training cycles to be reapplied. Transfer learning can drive efficiency savings by shortcutting model development. It can also be useful when data for the task that the model is being developed for is somewhat limited. But it’s important to note that the approach has limitations. Models that rely on transfer learning to gain generalized capabilities will likely require training on additional data in order to perform well in their domain of focus (See:Fine tuning) Validation loss is a number that tells you how well an AI model is learning during training — and lower is better. Researchers track it closely as a kind of real-time report card, using it to decide when to stop training, when to adjust hyperparameters, or whether to investigate a potential problem. One of the key concerns it helps flag is overfitting, a condition in which a model memorizes its training data rather than truly learning patterns it can generalize to new situations. Think of it as the difference between a student who genuinely understands the material and one who simply memorized last year’s exam — validation loss helps reveal which one your model is becoming. Weights are core to AI training, as they determine how much importance (or weight) is given to different features (or input variables) in the data used for training the system — thereby shaping the AI model’s output. Put another way, weights are numerical parameters that define what’s most salient in a dataset for the given training task. They achieve their function by applying multiplication to inputs. Model training typically begins with weights that are randomly assigned, but as the process unfolds, the weights adjust as the model seeks to arrive at an output that more closely matches the target. For example, an AI model for predicting housing prices that’s trained on historical real estate data for a target location could include weights for features such as the number of bedrooms and bathrooms, whether a property is detached or semi-detached, whether it has parking, a garage, and so on. Ultimately, the weights the model attaches to each of these inputs reflect how much they influence the value of a property, based on the given dataset. This article is updated regularly with new information.

14 hours ago

View

The browser wars aren’t about search anymore — here are the best alternatives to Chrome and Safari

The browser wars have entered a new phase this year: the fight isn’t just over search results anymore, it’s over which company’s AI gets to act on your behalf inside the browser itself. Google Chrome and Apple’s Safari still dominate the market overall, with Chrome’s edge coming largely from how aggressively it has woven generative AI into search. But 2026 has brought a wave of new entrants — from well-funded startups to Big Tech itself — all betting that the browser is about to become less like a window onto the web and more like an assistant that gets things done for you. Users looking for alternatives to Chrome and Safari can choose from a growing variety of browsers aimed at challenging the industry giants. To help navigate the competitive landscape, we’ve compiled an overview of some of the top alternative browsers available today. This includes browsers leveraging AI, open source browsers that promote customization and privacy, and “mindful browsers” — a new term that refers to browsers designed to enhance user well-being. Perplexity is the most recent startup in the space tolaunch an AI-powered web browser. CalledComet, the company’s new product acts as a chatbot-based search engine, and can perform actions like summarizing emails, browsing web pages, and performing tasks such as sending calendar invites. It’s currently only available to users with Perplexity’s $200/month Max plan, but there’s also a waitlist where people can sign up. The Browser Company, the startup behind the Arc browser,recently introducedDia, its AI-centric browser that looks similar to Google Chrome but with an AI chat tool. Currently available as aninvite-only beta, Dia is designed to help users navigate the web more easily. It’s able to look at every website that a user has visited and every website they’re logged into, enabling it to help you find information and perform tasks. For instance, Dia can provide information about the page a user is currently browsing, answer questions about a product, and summarize uploaded files. To get early access to Dia, users have to be an Arc member. Non-members can join the waitlist. Anotherrecent entryinto the AI agentic browser war is Opera’sNeon, which has contextual awareness and can do things like researching, shopping, and writing snippets of code. Notably, it can even perform tasks while the user is offline. Neon is currently available on macOS and Windows. The subscription costs $19.90 per month. OpenAI recently launched its AI-powered web browser, calledAtlas. The browser allows users to ask ChatGPT about search results and browse websites within the chatbot instead of being directed to outside links. There’s also an “agent mode” for users to ask ChatGPT to complete tasks on their behalf. Atlas was first rumored to launch inJuly; however, it only became available on macOS in October. It’s expected to arrive on Windows, iOS, and Android devices soon. Backed by Y Combinator,Asideis an upcoming AI-first, browser-native automation platform built to autonomously complete tasks, fill out forms, and manage data on behalf of users. The company describes the experience simply: “Give it your passwords, browsing history, and browser context.” Unlike traditional automation tools that rely on integrations, Aside operates directly within the browser itself, allowing it to work across Gmail, Notion, Slack, Figma, and banking platforms. Users can sign up for the waitlist ahead of launch. Jatterlaunched its AI-powered browser in June, giving users the ability to ask questions about any webpage, uncover relevant insights, and receive personalized recommendations based on their browsing activity. Additionally, Jatter offers an integrated Notes app, so it can learn from that content, summarize notes, and surface key details. Jatter is currently available on Mac, Windows, iOS, and Android devices. It’s free to use, but there’s also an optional subscription for $10 per month. Braveis among the more well-knownprivacy-first browsers, popular for its built-in ad and tracker blocking capabilities. It also has a gamified approach to browsing, rewarding users with its own cryptocurrency called Basic Attention Token (BAT). When users choose to opt in to view ads, supporting their favorite websites, they get a share of the ad revenue. Additional features include a VPN service,an AI assistant, anda video calling feature. DuckDuckGois anotherbrowserthat many people are probably already familiar with, thanks to its search engine by the same name. Launched in 2008, the company recently made significant investments in its browser to stay competitive byintroducing generative AI features, such as a chatbot. It alsoenhanced its scam blockerto detect a wider range of scams, including fake cryptocurrency exchanges, scareware tactics, and fraudulent e-commerce websites. In addition to blocking scams, DuckDuckGo prevents trackers and ads, and it doesn’t track user data, resulting in fewer pop-ups for users. Ladybird, led by GitHub co-founder and former CEO Chris Wanstrath, has an ambitious mission compared to other rivals: It aims to build an entirely new open source browser from scratch. This means it will not rely on code from existing browsers, a feat that has rarely been accomplished. Most alternative web browsers depend on the Chromium open source project maintained by Google, which is the most widely used base for many browsers. Like other privacy-focused browsers, Ladybird will offer features to minimize data collection, such as a built-in ad blocker and the ability to block third-party cookies. The browser has yet to be launched, with an alpha version scheduled for release in 2026 for early adopters, available on Linux and macOS. Vivaldiis a Chromium-basedbrowsercreated by one of the original developers of the Opera browser. Its biggest selling point is its customizable user interface, which allows users to change the appearance and enable or disable features. One unique feature is that the browser window changes color to match the website being viewed. Other key features include ad blocking, a password manager, no user data tracking, and productivity tools such as a calendar and notes. Operalaunchedthe Air browser in February, becoming one of the first mindfulness-themed browsers in the space. WhileOpera Airfunctions like a typical web browser, it includes unique features designed to support mental well-being. These features consist of break reminders and breathing exercises. Another feature, called “Boosts,” provides a selection of binaural beats to either help improve focus or relaxation. SigmaOSis a Mac-only browser featuring a workspace-style interface that emphasizes productivity. It displays tabs vertically, allowing users to treat them like a to-do list that can be marked as complete or snoozed for later. Users can create workspaces — essentially groups of tabs — to better organize different activities, such as separating work from entertainment. This Y Combinator-backed browser hasbeen aroundfor a few years now and has most recently begun introducing moreAI features,including the ability to summarize various elements of a web page, such as ratings, reviews, and prices. It also has anAI assistantthat can answer questions, translate text, and rewrite content. SigmaOS is free to use, but users who want more than three workspaces can subscribe to a plan for $8 per month, which provides unlimited workspaces. Zen Browseraims to create a “calmer internet” with its open source browser. Zen lets users organize tabs into Workspaces, and offers Split View to view two tabs side by side, among other productivity-focused features. Users can also enhance their browsing experience with community-made plug-ins and themes, such as a mod that makes the tab background transparent. This story has been updated after publication to include newly launched browsers.

18 hours ago

View

Why AI Startups Are Knocking on Corporate Doors Instead of VC Offices

MNCs are redesigning their accelerator programmes around business collaboration rather than investment.

22 hours ago

View

Submit your Tool

BestAITools.online is an AI Tools Directory helping individuals, businesses, and creators discover the best AI tools for writing, coding, design, productivity, and more.

Contact Promote Terms of Service Refund Policy Privacy Policy Email

AI’s promise to indie filmmakers: Faster, cheaper, lonelier

Latest AI News

In Swiggy's Developer Community, Some are Planning Parties, Other are Prepping Meals

The only AI glossary you’ll need this year

The browser wars aren’t about search anymore — here are the best alternatives to Chrome and Safari

Why AI Startups Are Knocking on Corporate Doors Instead of VC Offices

Quick Links

AI Tools by Task

Alternatives

Best AI Tools

Top AI Tools

Browse & Compare

Featured Lists

Discover