seen from Brazil
seen from United States
seen from Türkiye
seen from Sweden
seen from China
seen from United States
seen from China
seen from Canada
seen from United Kingdom
seen from United States
seen from Australia
seen from United States
seen from Hong Kong SAR China
seen from United States
seen from Malaysia
seen from United States
seen from United Kingdom
seen from China
seen from United States
seen from Germany
Am Arbeiten und wie der Name schon sagt am "Alexanderplatz" hehe #alexanderplatz #berlin #bigcity #world #models #actors #male #female #business #work #anzug #platz #germanmodels #worldmodels #internationalmodel #commercialshoot (hier: Berlin, Germany)
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch • No registration required • HD streaming
AI That Understands Physics: Why We Cannot Understand It Back
A Co-Written Essay with General Intuition's Pim DeWitte
This was a very long read, but I did read to the end (yep, got sucked down a rabbit hole) even though I didn't exactly understand every single concept that Pim DeWitte talks about here. I think many researchers struggle to explain their work to the average person, just because there is a baseline of knowledge/jargon that you need to know in that researcher's domain, and they often forget what normies don't know.
Anyway, I had a couple of reactions here. One is that it would seem like most of the work being done in world models is still based upon visual inputs, whether it is they are truly rendered into images or not as part of the simulation environment. To me, this would seem to limit the applicable use cases to ones where visual input is the only determinant of how a human would behave. And I'm still not sure how even a world model simulation will create exceptional conditions that it hasn't seen before. Take the example of a self-driving car, encountering a novel situation that doesn't appear in its training set (a schoolbus that's jack-knifed and is blocking the road, say). You can't simply dream up all the possible visual phenomena because that's an unbounded set. Yet a human driver would know exactly what to do in most emergent conditions. So I still don't get it.
I do agree that generative models (LLMs) are over-hyped in their capabilities (something Gary Marcus has been saying since at least 2019 and Yann LeCun is only getting credit for this because he's good at ripping off others' ideas and parroting them) and if we want to truly invest in simulating humans, then a more world model approach is needed. But again, I'm concerned that applicable use cases are far more narrow than the massive shovel-loads of venture capital would suggest.
Which brings me to my second point: These are clearly valuable angles of research, and the folks engaging in it seem very bright. Yet I feel like in the past, much of this research and experimentation would have happened in top-notch engineering schools, rather than in private industry. The trouble with trying to run a lab as a venture-backed business is that there is always going to be the pressure to commercialize before technology is ready, and also, the VCs have essentially set the expected value for the technology before it's proven. (Say you're a company like Fei-Fei Li's, who has raised about $1B -- that assumes the enterprise value of the company at exit is going to be north of $10B.) What if all of this world model research ends up being a dead-end? It's happened before in AI, many times (SVMs were supposed to be the future until they were not). Maybe all of this "experimentation is happening in private companies" is a function of institutional funding/grants collapsing, especially under the Trump Administration, but the problem is that the impacts of failure inside the university research system are relatively low (grants went to waste) versus in the private markets, where these $ are literally people's pension dollars being invested on extremely risky bets.
February 2026 Fashion cover, fashion magazine is Harper's Bazaar, Spain, cover fashion model is Adriana Lima, in "Eternal Flame", fashion photographer is David Roemer, one of their several covers. MUA: Frank B, hair styling by Shin Arima, stylist is Elizabeth Sulcer. General style adviser: Celia Cuervo. Acknowledgement TheFashionSpot. [At our BERLiNiB social media we post, as a rule of thumb, a pair from our OWN fashion magazine, which has only authentic non-ai photos, and a cover or so from ANOTHER good fashion magazine --and always with full and unambigious references. All the photos in the BERLiNiB magazine are of course entirely original and created by/for BERLiNiB--each team referenced; only at social media do we, sometimes, post news from other quality fashion magazines, and in doing this, we honor copyrights, which means that when we post from another magazine, we only do so for photos that are well represented in social media or at websites like models.com and thefashionspot.com, and always w/proper acknowledgements.]
In the measured cadences of DeepMind's latest, Genie 3, one glimpses the next chapter of world-modelling: environments that breathe, respond, persist under scrutiny. Not mere spectacle, but a simulation of consequence. For a nation long attuned to the subtle interplay of mind and machine, it offers a gentle reminder that progress arrives not with fanfare, but with incremental fidelity to reality itself.
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch • No registration required • HD streaming
Learn how AI uses world models to predict, plan, and improve itself on the path to superintelligence. https://www.gadgetaiworld.com/2025/09/superintelligence.html
Headed to Missouri for a agency photo shoot!! Booked a gig with my hom boi #WorldModels #musician #gig #evolusical #traveling #FollowtheJourney
