Something I find really interesting about @skyfallai is how they're bringing world models into actual engineering workflows. Their new model, SKY-001, can simulate designs, break complex objects down into individual components, and then reconstruct them in 3D. They're reporting 100x faster inference and 150x lower costs than GPT Astra for CAD tasks, which is pretty impressive. And looking at their roadmap, CAD seems to be just the beginning. They're exploring everything from PCB design and robotics to enterprise workflows. Really curious to see how far @spisallyouneed and the team can take this, especially at this pace.

Your model is already good, but what if it's working with the wrong context? You can give a model the best reasoning capabilities in the world, but if the information it's working with is wrong, the answer will be wrong too. That's why I find what @mynameisyahia is building with Context really interesting. I feel like the next big step towards better AI agents isn't just improving models, but building the infrastructure that makes the internet more accessible to them. Better context means better answers.

I've gotten pretty good at figuring out SEO, but AEO is a completely different game. I've struggled with it myself, and the fact that people are paying thousands of dollars just to figure out AEO says a lot about how difficult it still is. Really cool to see Treg launch something that tackles this. I'll definitely be using it for my own products. And the crazy part? It costs just 1 cent.

One thing I really wanted to fix with lattedit was how painful it is to actually work on LaTeX projects with other people. So we built real-time collaboration, inline comments, and an AI agent powered by Sarvam 105B that can directly edit your LaTeX. You can now also convert PDFs back into editable LaTeX while preserving the layout, import your Overleaf projects, and start with ready-made templates. And yes, everything is free. lattedit.xyz

Such a cool launch video, and most importantly, my favorite creator platform is now on LinkedIn!

Our biggest launch at Solacc is here! We're introducing an AI Interviewer that conducts mock interviews tailored to your resume and job description, helping you prepare for the questions that actually matter. We're also launching Real Interview Experiences, where you can share questions from your interviews, explore others' experiences, and learn what companies are really asking. And the best part? You can try your first AI interview for FREE right now! Practice with AI. Learn from real interviews. Ace your next one.

launching lattedit.xyz A calm little LaTeX editor for your browser, Free, no account, works on your phone. LaTeX, freshly compiled.

About 1.5 years ago, I published a paper where I found early traces of visual reasoning in Gemini. Ever since, I've wanted to explore this idea further. The thought of AI actually being able to think visually fascinated me, and I genuinely believed there was something special here. Seeing Glimpse bring this idea to life today feels surreal. Something I wanted to work on for so long is actually becoming a reality. This is just awesome.

Who owns something created with AI? I make a lot of AI-generated launch videos, and this is something I've wondered about myself. VN Studio takes an interesting step toward answering that question by tracking and documenting the human creative decisions behind every edit. I think this is the right direction for copyright in the age of AI.

Back in the day when you didn't have AI, you just had to Google search and find stock photos but you would never find the one that you would actually use. Only if there was something to edit in the missing piece 😞 Envato lowkey solves that problem. Maybe you need a picture of a zebra standing near a giraffe. Just put the giraffe now using AI inside Envato.

We have a similar thing here in India for this, but it is actually done through voice. That is text and doing it through text makes it much more sensible. Now that call center work can be completely given to AI agents and they'll actually do it very seamlessly. Good work from Bland.

One of my biggest frustrations with using Codex or Claude is having to log in to different services over and over again just so the agent can get things done. And honestly, I don't trust these agents enough to hand over my credentials either. They can do pretty much anything with that access, right? The idea of an AI agent being able to use my accounts without ever knowing my login credentials is genuinely exciting. It solves a very real problem I've run into countless times. Cool stuff, man.

I once went to an IIT where they were still processing payments through Excel. Just like that, many other companies still do that. So many things scattered across so many different files. Then I tried out Proto. They are actually combining all the sheets, and AI agents are running the whole operations. This is a pretty interesting use case for AI beyond simple chatbots.

One thing that really stood out to me from this paper and this research work is that a previous image jailbreak method reported 31% success when evaluated at human-calibrated criteria. That just dropped to 0.3%, while Rise managed up to 13% just by making new attack strategies instead of individual prompts. Really makes a question: How reliable are judges in AI safety benchmarks?

one of the craziest things i liked from this launch is this hotel rating i had once booked a 5 star hotel in hyd and it would have good ratings everywhere but it's wellness and amenities was so bad! only if i had a tool like odessia to help me with that :( usually those metrics are hidden deep in reviews that we never find em.