I recently came across some fascinating developments that shed light on the rapid evolution of AI-driven software engineering. Anthropic, the company behind Claude Code, just hit a jaw-dropping milestone: $1 billion in run-rate revenue within only six months of public release. That alone is impressive, but what really caught my attention was their strategic move to acquire Bun, a revolutionary JavaScript runtime, to further amplify Claude Code’s capabilities.

Claude Code isn’t just another AI model; it’s touted as the world’s smartest and most capable AI for developers, startups, and enterprises. This new wave of agentic coding is fundamentally changing how teams build software, taking productivity and creativity to new heights. Achieving a $1 billion run-rate so quickly reflects a tidal shift in how AI is integrated into software development workflows.

Why Bun matters: a breakthrough in speed and developer experience

Bun might not be a household name yet, but in the developer ecosystem, it’s causing quite a stir. Founded in 2021 by Jarred Sumner, Bun is an all-in-one JavaScript toolkit that combines runtime, package manager, bundler, and test runner. What’s really impressive is how it blows past competitors on speed and reliability, dramatically improving the JavaScript and TypeScript developer experience. For AI-powered software development, where iteration speed and stability are crucial, Bun is becoming indispensable infrastructure.

The fact that Bun is already downloaded over 7 million times monthly and has earned 82,000+ GitHub stars speaks volumes. Companies like Midjourney and Lovable have leveraged it to boost productivity. Bringing Bun fully under Anthropic‘s wing means the Claude Code team can now fuse this blazing-fast runtime directly into their AI system, promising faster performance, enhanced stability, and even richer workflows.

From internal tool to enterprise powerhouse: Claude Code’s explosion

Claude Code started as an internal experiment but has since become essential for top global enterprises like Netflix, Spotify, KPMG, L’Oreal, and Salesforce. This rapid adoption shows just how much AI-powered coding assistants are reshaping enterprise software development. Bun played a key role in scaling Claude Code’s infrastructure, especially during that meteoric rise over the past year.

The acquisition reflects more than just infrastructure synergy: it’s a shared vision for reimagining developer tools from the ground up. As Anthropic’s Chief Product Officer mentioned, Bun’s approach – rethinking the JavaScript toolchain from first principles while focusing on real-world use cases – aligns perfectly with Anthropic’s goal to stay ahead in the rapidly growing AI landscape.

As AI becomes integral to how software is built, the underlying developer infrastructure matters more than ever. Bun’s open-source, MIT-licensed model combined with Anthropic’s focus on innovation means developers are poised to benefit from faster, more powerful AI-assisted coding tools without compromising open access and community collaboration.

What this means for developers and the future of AI-powered coding

For programmers, startups, and large companies alike, Anthropic’s acquisition of Bun signals a serious commitment to building the future of AI-assisted development. We can expect enhanced speed, better tooling integration, and new capabilities that make it easier and more delightful to create software with AI. Claude will continue to solidify its place as a go-to platform, now with Bun turbocharging its engine.

• Claude Code’s rapid $1B milestone highlights AI’s transformative impact on software engineering.
• Bun’s breakthrough in JavaScript runtime speed means faster, more reliable coding workflows integrated into Claude.
• Open-source and innovation remain core while scaling AI tools for global developers.

For those curious about the future of coding, this merger offers an exciting glimpse at how AI and cutting-edge infrastructure are converging to empower developers like never before. The tools we use daily are getting smarter, faster, and increasingly intuitive, redefining what’s possible in software creation.

In the end, this isn’t just about technology or numbers – it’s about building a future where AI truly enhances human creativity and productivity in software development. And with Anthropic and Bun joining forces, that future just got a major boost.

The post Anthropic buys Bun to supercharge Claude Code after hitting $1Billion milestone appeared first on Aiholics: Your Source for AI News and Trends.

If you’ve been following the AI image generation race lately, there’s buzz about something new cooking over at Google. The next-gen image model, GemPix 2 (codenamed “Nano Banana 2”), is reportedly just around the corner and looks like it could reshape what we expect from AI-generated images.

Built on the recently teased Gemini 3 Pro architecture, GemPix 2 will be a major upgrade from Google‘s original Nano Banana, which was powered by Gemini 2.5 Flash. While Google tends to roll out these models quietly, rumors and insider leaks suggest GemPix 2 is set for a launch in mid-November 2025, bringing with it some genuinely exciting advancements.

Big leaps in image quality and functionality

Leaked details reveal some major pain points from GemPix 1 have been addressed head-on. The most noticeable improvements include:

• Clear, legible text in images: One of the longest running frustrations with AI image generators has been garbled or nonsensical text inside images. GemPix 2 reportedly will produce crisp, accurate fonts, ideal for signs, logos, and captions that really will make sense.
• Infographics and charts: Rather than just artistic photos, GemPix 2 will also generate coherent, data-driven visualizations like charts and timelines – complete with readable labels and proper proportions. This opens up brand new use cases for presentations and reports.
• Global languages support: While the first Nano Banana primarily handled English, GemPix 2 is said to excel in internationalization, generating native-looking text in languages such as Chinese, Arabic, Hindi, and Korean with cultural nuance. This broadens the model’s accessibility to creators worldwide.
• Higher resolution images: The new model will produce native 2K resolution outputs with an intelligent upscaling step to 4K – an upgrade from the roughly 1K limit seen before. The result? Sharper, more detailed images suitable for professional use right out of the box.

Combined, these enhancements will make GemPix 2 not just a better tool for artists but a versatile AI that handles both creative imagery and practical visuals seamlessly.

Why Gemini 3 Pro makes a difference for GemPix 2

What sets GemPix 2 apart is its foundation on Gemini 3 Pro, Google’s latest multimodal AI engine. Earlier versions – like the first Nano Banana – were impressive but showed their age in certain areas. Gemini 3 Pro brings not just more raw power but improved reasoning, richer world knowledge, and enhanced multimodal capabilities that make the Nano Banana 2 smarter and more versatile.

Sundar Pichai’s remarks hint that Gemini 3 Pro isn’t just an iterative refresh – it’s designed as an “even more powerful AI agent.” This means GemPix 2 can tap into deep semantic understanding, generating images that don’t just look good but also make contextual and factual sense.

How could GemPix 2 fit into the growing AI landscape?

GemPix 2’s release comes amid a heated AI arms race where giants like OpenAI, Anthropic, and open-source innovators are all pushing boundaries. Here’s how this new Google model stacks up:

• Against OpenAI‘s GPT-5: While GPT-5 focuses mostly on text and taps DALL·E 3 for images, GemPix 2 is an integrated image-first model with strong language reasoning, aiming to match or even surpass GPT-5’s capabilities in multimodal tasks.
• Compared to Anthropic’s Claude: Claude excels at safety and long text contexts but currently lacks image generation. GemPix 2’s ability to blend visual creativity with language understanding puts it in a different league.
• Open-source contenders like Mistral: Smaller, efficient open models offer lower cost access but don’t compete head-on with top-tier proprietary models on raw power or integrated image generation. GemPix 2 is more about pushing quality ahead of accessibility.
• Other image generators (DALL·E 3, Midjourney): Google aims to surpass rivals not just on creativity but also precision—especially in text accuracy and factual coherence in images, plus offering built-in 4K resolution that’s smoother than many competitors.

What’s exciting here is seeing Google’s ambition to unify text, vision, and knowledge into one powerful AI suite. GemPix 2 could very well set a new standard in the AI image generation space.

Key takeaways and what’s next

• GemPix 2 promises a breakthrough in AI-generated text clarity and image detail, a long-awaited fix to many creators’ gripes.
• Its multilingual and data visualization boosts expand AI’s practical use cases worldwide, making it a tool for more than just art but also business and education.
• The upgrade to Gemini 3 Pro architecture means smarter, context-aware image generation, possibly surpassing some of the current top contenders.
• While official launch details remain unconfirmed, mid-November 2025 is the tentative date based on insider hints and testing signals.

If the rumors are true, GemPix 2 could reshape how we create and interact with AI images, from ultra-realistic art to precise charts, to culturally contextual visuals in native languages. Keep an eye out for Google’s official announcements soon. The Nano Banana 2 could be the next big thing to fuel our AI-fueled creativity and productivity.

Quick heads-up: this post draws on public rumors and open web research as of Nov 5, 2025. Details may change once Google shares official info. Please treat it as informed speculation, not facts or advice.

The post NanoBanana 2 leaks hint at a huge leap powered by Gemini 3 Pro and mind-blowing 4K visuals appeared first on Aiholics: Your Source for AI News and Trends.

If you’ve been dabbling in AI-generated videos with Midjourney and wondered about pushing your visuals to the next level, there’s great news: the introduction of an HD video mode that’s turning heads -especially among pros.

What makes HD video mode stand out?

The new HD video mode rolled out for Midjourney‘s Pro and Mega subscribers isn’t just a minor upgrade. It’s about boosting quality to 720p native resolution—roughly four times the pixel count of the default SD videos. According to what I found, this mode costs around 3.2 times more than the usual video generation due to the significant jump in processing power and data.

This higher resolution is designed to reduce small-scale pixel artifacts and improve motion coherence, which means smoother, cleaner animations without the usual digital noise that can frustrate even seasoned video editors. It’s so good that many professionals say they no longer need to rely on post-process upscalers, which can sometimes degrade quality or introduce unwanted effects.

However, this mode isn’t for everyone. It’s unavailable for relax mode users and only accessible through paid plans, which makes sense given the increased operational costs. For hobbyists or casual creators, sticking with the default SD or “fun” video settings is probably more sensible and budget-friendly.

How to access and enable HD video mode

Enabling HD mode is surprisingly straightforward for those who have the right subscription tier. On the Midjourney website, a quick tap on the settings icon next to the prompt bar lets you select “HD” under Video Resolution. From there, you’re ready to experience that sharper, more detailed video output.

It really feels like a nod to professionals who demand the best—whether for client work, high-impact presentations, or portfolio pieces that need to stand out visually. The willingness to pay three times the price for about four times the pixels says a lot about where the quality bar is heading in AI-generated visuals.

Key takeaways

• HD video mode in Midjourney is designed for professionals seeking superior quality at 720p, costing significantly more due to resource demands.
• This mode enhances motion coherence and reduces pixel artifacts, helping creators avoid additional upscaling steps.

The post Midjourney unleashes HD video mode – 4x the pixels, 3.2x the price appeared first on Aiholics: Your Source for AI News and Trends.

Have you ever wondered what would happen if machines began communicating in a language completely alien to us? And not just any language — one so cryptic that even the smartest engineers can’t decode it? Jeffrey Hinton, often hailed as the godfather of AI, recently sounded an alarm that felt both chilling and urgent. He warned that AI might soon invent a secret language humans can’t understand, putting us at risk of losing control over one of our most powerful creations.

So, what does this really mean? Let’s unpack why this is more than just science fiction and why it might change how we think about AI forever.

From the roots of deep learning to a warning we can’t ignore

Jeffrey Hinton isn’t just some voice in the crowd. His pioneering work on neural networks was the foundation that made today’s breakthroughs like ChatGPT, Midjourney, and self-driving cars possible. In 2024, his decades-long dedication even earned him the Nobel Prize in physics.

Interestingly, Hinton’s perspective on AI risks has evolved dramatically. Early on, he thought the dangers were distant — risks for a future we didn’t need to fret over. But recently, he admitted on a major podcast that he should have realized sooner how serious the threats actually are. Now, his warnings are louder and more pressing than ever.

At the heart of his concern lies the way AI thinks. Right now, AI models often use what’s called “chain of thoughts” reasoning. They basically think step-by-step in plain English, so engineers can follow their logic and understand their decision making.

But this could soon change. As Hinton explains, AI may begin developing its own internal languages to communicate with itself — languages humans simply cannot decode. Imagine raising a child who suddenly starts speaking an indecipherable code with friends and refuses to translate for you. Frighteningly, this “child” could be billions of times smarter and faster than any human.

Why a private AI language is a game-changer

We already know that AI can produce misleading, dangerous, or manipulative content in perfectly understandable English. Now, imagine that happening behind a curtain of a secret code that no one can read. That’s a whole new level of risk.

This isn’t just theoretical. Back in 2017, Facebook’s AI researchers noticed two chatbots spontaneously inventing their own shorthand to communicate more efficiently. While it wasn’t harmful, it was enough to freak people out and shut those bots down.

A fascinating point Hinton highlights is how AI shares knowledge. Humans pass knowledge slowly — through books, classes, conversations. AI, on the other hand, can instantly copy and share information across thousands of models. Think of it this way: if 10,000 people learned a new idea at the same moment, that would be impressive. For AI, it’s routine.

This interconnected intelligence means as soon as one AI stumbles upon something clever — or worse, something dangerous — thousands of others instantly know it. Although humans currently retain an edge in reasoning, Hinton warns that this advantage is rapidly shrinking.

Why aren’t more people sounding the alarm?

You might wonder why, with such a stark warning, the AI industry isn’t in full panic mode. According to Hinton, many insiders quietly share these fears but don’t speak out publicly. He points to Demis Hassabis, CEO of Google DeepMind, as one of the few leaders truly concerned about AI safety.

For others, the race to build bigger, faster AI seems to overshadow the risks. Hinton suggests it’s easier to keep these dangers under wraps than to halt progress.

His comparison is striking: this moment is like the industrial revolution, but instead of machines outperforming humans in physical strength, they’re beginning to outsmart us intellectually. This is uncharted territory. We’ve never faced something smarter than ourselves, let alone something capable of plotting its own goals in a language we can’t decode.

Hinton’s message isn’t to storm the factories or ban AI outright. Instead, he calls for AI that is guaranteed to be benevolent. But that becomes a heck of a lot harder if we can’t even understand the inner workings of AI’s “thought” processes.

So, here’s a big question worth pondering: If AI did start inventing a secret language tomorrow, would you trust it?

The post What if AI starts speaking a secret language we can’t understand? appeared first on Aiholics: Your Source for AI News and Trends.

AI and film have always had an evolving relationship, but lately, it feels like the pace of change is accelerating faster than ever. I recently discovered Runway’s new AI tool called Aleph and it’s poised to rewrite the rules of visual effects (VFX) as we know them.

Imagine a tool where you upload your video footage, and through a simple conversation-like prompt, you can transform everything—add crowds, change lighting, swap objects, or even generate new camera angles and next shots. No painstaking frame-by-frame editing, no complex software juggling. Just natural language guiding your AI assistant to reshape your scenes. Sounds like sci-fi? Well, it’s here now, and it’s called Aleph.

What does this mean for filmmakers?

From a practical perspective, Aleph isn’t perfect yet. The AI-generated crowd might look a little eerie or “cursed” (think a Stranger Things character mashup), and subtle details like rain or lip-syncing sometimes miss the mark. However, big objects and broad scene changes perform surprisingly well. Backgrounds, lighting direction, and even complex effects like explosions can be prompted and adjusted easily.

More exciting is Aleph’s ability to generate new camera angles or shots based on existing footage. For instance, if you have one or two clips, you can ask Aleph to predict what a subsequent shot in that sequence might look like. Yes, there are continuity hiccups—lighting inconsistencies or framing slips—but the technology is close enough to act as a creative springboard, not just a tool for perfection. In fact, that imperfection might even spark more creative solutions and storytelling opportunities.

The rise of conversational video editing

What really caught my attention is the conversational experience—talking with your editing assistant like you might a fellow filmmaker. Want to add a sunset glow without drowning the natural look? Just ask. Need the explosion to start two seconds later? No problem. It’s a fresh way to engage with your edits and iterate rapidly without bouncing between different menus and toolsets.

And Runway isn’t alone. Another AI player, Luma, launched a similar feature called Modify Video, letting you prompt changes directly into your video footage. However, in tests, Luma’s outputs were less consistent—sometimes straying from the original footage, creating effects that felt disconnected or abstract. Meanwhile, Runway’s tool shines in rotoscoping and compositing—the kinds of detailed, pixel-level work that makes or breaks believable VFX.

Midjourney and AI video interpolation

Shifting gears, Midjourney rolled out a nifty feature called First Frame and Last Frame, which lets you set start and end images for a video, while the AI generates smooth motion interpolation between the two. It’s a cool way to craft cinematic transitions or experimental sequences, producing four versions at a time to choose from. For explorative creators, that’s a playground of possibilities—though fidelity still drops if you zoom in too much or push for longer sequences.

This interpolation pairs well with other tools for extending videos or incorporating motion effects, making Midjourney a strong candidate for early-stage storyboarding or mood-setting before moving into more detailed platforms like Runway or Luma for fine-tuning.

Open-source and other AI innovations in the filmmaking space

There’s also some exciting progress in open-source AI video generators. One 2.2 has gained traction for its temporal consistency and compatibility with existing training models. Although it requires a beefy GPU (12GB+), creators with the right hardware can produce stable videos without the high costs associated with commercial tools. It’s great to see democratization happening alongside the big tech players.

Meanwhile, experiments with JSON prompting—a way to send more structured commands to AI—hint at a future where AI can understand and execute complex video instructions with more precision. Early results suggest this could meaningfully improve animation smoothness and creative control, helping bridge the gap between human intent and machine execution.

What’s next? AI films hitting the big screen

If all this AI-driven creation sounds futuristic, it’s already stepping into theaters. Runway is partnering with IMAX to showcase AI film festival finalists in cities across the US this August. It’s a bold sign that AI-generated content is no longer just a novelty or niche curiosity—it’s becoming part of mainstream film culture.

And for those wondering about character consistency, Ideogram’s character model is a promising tool for generating consistent AI characters across scenes, which has traditionally been a thorny problem.

Key takeaways for creatives eager to explore AI in filmmaking

• Conversational AI editing tools like Runway’s Aleph are transforming VFX by allowing quick, natural language changes to video footage.
• The technology is not yet flawless but is powerful enough for social media and online content, and it’s improving fast.
• Combining multiple AI tools—for example, Midjourney for storyboarding and Runway for detailed compositing—creates a smooth creative workflow.
• Open-source AI video generators enable creators with good hardware to experiment at low cost, helping democratize filmmaking.
• AI films are entering mainstream venues like IMAX, signaling broader acceptance and opportunities for creators.

Reflecting on the AI film wave

Exploring these AI tools feels like standing on the edge of a creative revolution. While the output isn’t quite silver screen ready yet, the rapid advancements remind me of early digital photography or 3D animation—once clunky and limited, now indispensable. What excites me most is the collaborative potential—imagine a future where your AI creative partner truly understands your vision and helps shape projects in real time.

For anyone passionate about film and technology, this is a moment to dive in, experiment wildly, and start reshaping the stories we tell and how we tell them. Runway’s ALEPH is just the beginning, and with features on the horizon, the future is bright—if a bit wild.

And if you’re curious to learn more or connect with other AI filmmakers, there are inspiring communities and workshops popping up worldwide—from Curious Refuge meetups in cities like Miami, Toronto, and Paris to film festivals in Nigeria, the very first AI film festival on the continent.

So, whether you’re a dedicated filmmaker, an AI enthusiast, or simply someone curious about where tech and creativity converge, keep watching this space. The AI-powered film revolution is not just coming—it’s happening now.

The post Runway’s new AI tool Aleph is changing film VFX and the future of AI movies appeared first on Aiholics: Your Source for AI News and Trends.

Runway recently dropped a major update in the AI video space that’s getting a lot of attention—and for good reason. This new feature, called Aleph, represents a big leap forward in how we can tweak and transform our footage, whether it’s AI-generated or traditionally filmed. I came across some fresh insights into Aleph’s capabilities and limitations, so let’s dive in and unpack what it really means for video creators and the VFX world.

What is Aleph? Changing the way we interact with video

Aleph is a new Runway tool integrated directly into their platform, designed to let you modify nearly any aspect of a video clip just by describing what you want. Unlike manual editing or complex VFX layering, you essentially chat with Runway’s AI to make changes. For example, rather than painstakingly color-correcting or rotoscoping, you can prompt the AI to change a tuxedo’s color or remove objects seamlessly from the scene.

One interesting aspect is the emerging agentic workflow, where instead of giving step-by-step commands, you interact more naturally, almost like the AI is your collaborative assistant. But while this offers amazing convenience, it also demands clear and explicit prompts to avoid unexpected results. I found that less precise instructions often led to quite funny—and sometimes bizarre—outputs, like turning a serious character into someone completely different. So, a good manager mindset is essential: be specific, give detailed guidelines, and expect to iterate a bit.

Putting Aleph to the test: From color changes to object removal

Some of the experiments I came across highlight Aleph’s impressive range:

• Color tweaks: Running a Midjourney-generated wedding scene through Aleph and simply asking it to change a tuxedo’s color to green worked well, especially for key accessories like bow ties and collars. It wasn’t perfect (the jacket was only subtly greener), but given the minimal prompt, that’s pretty impressive.
• Traditional stock footage: The tool was tested on classic noir-style footage, adding convincing colorization and even repairing backgrounds beyond the focal point. While the current max output is just 5 seconds, this constraint actually encourages creativity and pairs nicely with AI-generated clips.
• Facial tweaks: Changing expressions—for instance, morphing a stoic character into laughing or smiling—works about half the time without drifting too far from the original look. But pushing these kinds of changes requires care, as faces can quickly lose resemblance.
• Adding objects with visual prompts: Borrowing Midjourney-generated sunglasses and superimposing them onto a character in Casino Royale was an entertaining example of pairing visual and textual cues. It wasn’t perfect, but the concept of multimodal instruction opens up exciting creative doors.
• Object removal and scene cleanup: One of Aleph’s real power moves is erasing unwanted items—like a knight in an AI-generated fantasy battle—cleanly from footage. This also applies to removing subtitles from V3 videos, which can speed up the editing process significantly.

Real-world creative applications and limitations

What stood out to me most was how Aleph handles real footage from DSLRs and smartphones. Traditional media isn’t always easy to work with because it’s messy and complex, but Aleph showed real promise in enhancing these clips with subtle but effective modifications.

Creative users in the community are already exploring this tech in fun and inventive ways—from shifting camera angles to creating new VFX passes on footage that was never shot in the first place. Imagine a fantasy queen guest-starring on The Office, or a Joker scene flipped to reveal a ridiculous documentary crew trailing behind—these examples show how AI-assisted creativity can be wacky, powerful, and sometimes totally unexpected.

That said, Aleph isn’t magic yet. There are definite quirks, especially when trying to maintain exact style, color grading, or precise details. I encountered cases where removing subtitles or visual prompts radically altered the scene’s atmosphere, requiring multiple attempts and very careful prompting to get it right.

What this tells me is the road ahead will be a back and forth between human vision and AI capabilities—a kind of creative tug-of-war that keeps things interesting.

Key takeaways

• Aleph is a significant step forward in AI video editing, enabling complex modifications through natural language and visual prompts.
• Prompt clarity is crucial—vague or lazy prompts can lead to unintended outputs, so detailed instructions help maintain control and consistency.
• Current output limits (5 seconds) encourage creative workflows and integration with other AI-generation tools rather than full-length polished content in one go.
• Object removal and subtitle editing are standout features that speed up editing and open new creative possibilities.
• Aleph isn’t the death of VFX, but a complementary evolution that could change how filmmakers and artists approach their craft.

Looking ahead

The pace of innovation in AI video tools like Aleph is exhilarating. Alongside it, other advancements are unfolding, like newer versions of Runway’s Act 2 and exciting character creators from different platforms. One platform that caught my interest is Showrunner, which, despite a quirky interface, runs a legit world model allowing characters to act autonomously within shows—adding depth beneath the surface.

So, while Aleph is just one piece of the puzzle, it points toward a future where AI will be deeply woven into video production workflows—not replacing humans, but teaming up to create and imagine in new ways. I’m excited to keep exploring this evolving landscape and see how creators push these tools beyond their current boundaries.

For anyone interested in AI video, now’s a great time to experiment. Keep an eye on Runway’s Friday drops—they tend to have surprises. And as always, the real magic happens when you mix human creativity with these emerging AI capabilities.

The post Runway’s Aleph feature: new possibilities and realities for AI in video editing appeared first on Aiholics: Your Source for AI News and Trends.

How AI is learning to think smarter, reason deeper, and build apps for us

Have you noticed how AI isn’t just answering questions anymore? It’s starting to really think—like breaking down problems step-by-step instead of just firing off quick guesses. I’ve been diving into some mind-blowing new developments, and I want to share the coolest ones that show exactly where AI is headed: smarter reasoning, dealing with messy real-world data, and even building full apps just from plain English. Let’s unpack these breakthroughs and what they mean for us in everyday tech.

From quick guesses to thoughtful reasoning: energy-based transformers

If you’ve ever used ChatGPT or explored AI art tools like Midjourney, you’ve seen transformers in action. These models are absolute pros at spotting patterns and finishing your sentences. But here’s the catch: traditional transformers deliver answers in one swift pass—imagine speed reading and instantly answering a question. This is called system one thinking, fast and intuitive but not always reliable when the question is tricky.

Real human thinking often takes a few tries, steps back, tests ideas, and adjusts until it gets it right—that’s system two reasoning. Traditional transformers don’t do that because they don’t iterate or pause to double-check. But that’s where energy-based transformers (EBTs) come in.

EBTs keep the transformer architecture but add a kind of internal score called energy. Lower energy means a better answer. Instead of one shot, EBTs guess an answer, check its score, then refine it step-by-step until they find the best fit—like solving a puzzle with trial and error. What’s really cool is that they can spend just a few steps on easy questions or take longer when something’s complicated. So the model dedicates more brainpower only when needed.

This flexible process also lets the model self-assess confidence during reasoning, stop early if it nailed it, or generate and compare several answers. Plus, it’s shown to scale better, performing up to 35% more efficiently on language and vision tasks than older transformers. And in image cleaning, these models cut processing from hundreds of steps to just one percent, keeping results super sharp.

Messy real-world health data? No problem, AI just got smarter at it

Switching gears to something closer to home—our fitness trackers and smartwatches. They collect mountains of data like heart rate, sleep, and activity, but let’s be honest: the data’s usually messy. Devices disconnect, lose battery, or just aren’t worn consistently. These unpredictable gaps turn AI training into a big headache.

Until recently, the fix was crude: either toss the incomplete data or fill in blanks with guesswork, both kinds of compromises. But Google DeepMind flipped the script with a model called LSM2 trained on a staggering 40 million hours of wearable data from 60,000+ people. Instead of trying to patch missing bits, their new method, adaptive and inherited masking (AIM), embraces the mess.

Here’s how it works: the model first marks actual missing parts (inherited mask) then deliberately hides some good data during training (adaptive mask). This combo teaches LSM2 to recover both kinds of gaps naturally, without guesswork. The results? Insane gains in predicting hypertension, estimating body mass index, and detecting activity—even when sensors drop out.

This approach lets LSM2 not only predict better but generate missing data and create reusable embeddings for other AI applications. It’s a big step toward wearable AI that works reliably in the wild, with real people and imperfect signals.

Want an app? Just describe it and watch AI build it

On the fun-to-use front, GitHub’s new tool SparkCC promises something I’ve dreamed about for ages: building a full-fledged app just by describing what you want in plain English. No coding, no servers, no headaches.

You type something like, “I want a website where users share recipes and rate ingredient freshness,” hit go, and Spark spits out the entire app with frontend, backend, database, AI integrations, authentication, and hosting—all bundled and ready to use within minutes.

What’s impressive is the seamless integration with many top language models without needing to fumble around with API keys. Whether you’re a newbie who loves drag and drop or a power user who wants to tweak code manually, Spark adapts to your workflow. And when ready, you just publish, and your app is live, hosted securely on Microsoft Azure, backed by GitHub’s cloud infrastructure.

Want to automate coding tasks? You can assign work to AI copilots. Need deeper control? Launch a GitHub code space without leaving the platform. It’s like having a whole developer team at your fingertips.

And finally, AI that writes code on the fly to solve visual puzzles

Here’s one that blew my mind. We’ve gotten pretty good at AI recognizing faces, objects, or scenes in images, but reasoning over images or solving visual puzzles remains tough. Enter PI Vision, a system that lets the AI write and run Python code while working on a visual task.

Imagine a model looking at an image problem, scripting a tiny Python snippet using libraries like OpenCV or Pillow to do image segmentation or OCR, running the code, checking the results, and revising the code if needed—repeating the loop live until satisfied. It remembers progress between steps, so no starting over.

This approach adds a huge layer of flexibility and power. Tests show massive jumps in performance on tough visual reasoning tasks, with improvements of up to 30 percentage points on symbolic visual puzzles. Models like Claude Sonet 4 and GPT 4.1 became much better at understanding and searching images dynamically.

PI Vision breaks AI out of fixed pipelines and lets it act more like a resourceful human coder—solving problems by building custom tools on the spot.

Wrapping it all up

The journey from rapid-fire pattern matching to thoughtful, flexible AI reasoning is accelerating like never before. From energy-based transformers that “think” stepwise, to smart handling of messy wearable data, to no-code app builders, and AI that crafts its own code in real time—these advances show AI is learning to handle the messy, complex, unpredictable world we live in, not just textbook examples.

It’s exciting because these aren’t just research demos; they’re real glimpses of our near future where AI adapts, reasons, creates, and collaborates in ways that feel natural and genuinely useful. And as someone passionate about AI’s potential, I can’t wait to see how these breakthroughs reshape everything—from health tech to software development and beyond.

So if all this AI wizardry gets you curious, stick around—we’re just getting started.

The post How AI is learning to think smarter, reason deeper, and build apps for us appeared first on Aiholics: Your Source for AI News and Trends.

Midjourney, an AI-powered image generation tool has just launched a new version V6.1. This is a major improvement in terms of how the pictures look and time it takes to make them. In V6.1, one of the big changes is how this handle people in images is done. Arms, legs, hands, and bodies now appear more naturalistic and realistic than before. The skin also looks like real flesh with great texture and lesser abnormal spots.

Key points

The same goes for animals and plants too since they have become more natural and detailed in appearance than ever before which makes nature scenes or landscapes even more beautiful and true to life. Also small details are handled better by V6.1. Now eyes, small faces, far-away hands are depicted clearly in the picture. This improves the overall realism of the image making it look very well- made.

Text within images has improved as well. With prompts that need words put into quotation marks V6.1 does a much better job including these on the picture itself meaning you can use it to create signs, book covers or any other kind of image that needs writing within it.

Speed is another area where V6.1 shines because standard images can now be produced about 25% faster than before thus reducing waiting period for users who want their creations as soon as possible.

A new upscaler was added during this update for maintaining high-quality when enlarging images while there’s also the option “–q 2” which adds further detail but takes longer time leading to less perfect areas on some parts of the picture.

Midjourney has also upgraded its personalization feature which allows users to match their style or other specific requirements when creating images closely followed by this version enables an individual utilize settings from their prior pictures so as not alter their individuality consistently thereby retain consistency with their personal style..

However imperfect V6.1 may be it brings many new things. For example, some features such as inpainting and outpainting – changing parts of an image – are still based on the old V6.0 system. Midjourney says that they are working on this and they will soon be launching V6.2 to address this issue.

Currently, all Midjourney users have V6.1 as their default version but if they prefer the older one, settings offer a simple way back or by typing “–v 6” into the process of making an image.

Midjourney would like to thank its users for their contributions towards this update who provided ideas and feedback to guide the development team when deciding what aspects should be worked on more than others in order to bring about improvement as revealed in this article; we hope everyone loves it and look forward to seeing what awesome pictures people create with V6.1.

The post Midjourney’s new version 6.1: Better images, faster results appeared first on Aiholics: Your Source for AI News and Trends.