[ Removed by Reddit ]
[ Removed by Reddit on account of violating the content policy. ]
[ Removed by Reddit on account of violating the content policy. ]
Flew the Avata360 forward slowly through this gorge, maybe 120-150m up, roughly level with the ridgeline on either side. Weather wasn't great but I just sent it straight ahead into the clouds. Overall the Avata360 is easy to get comfortable with, but I'd say get a few flights in somewhere open before you take it to a spot like this. Wind in the mountains is a different thing and nobody wants your first real flight to be somewhere you can't just land it if something feels off.
I’m putting together a DIY cold plunge and I’m starting to think I really underestimated the chiller part. At first I figured an aquarium chiller would probably be fine, but now I’m not so sure. My setup will likely be in a garage, and I keep seeing people say smaller chillers take forever to cool a stock tank down, especially once the room gets warm. Getting the water cold one time is one thing, but keeping it cold after someone actually gets in seems like the harder part
For anyone running a DIY setup, how long does yours take to get from tap water temp down to around 45°F? What size tub are you using, is it insulated, and what HP chiller did you end up with? I’ve been looking at stronger options like the IceDragon Core because I don’t want to buy something that has to run nonstop just to keep up, but I’m still trying to figure out what’s actually necessary and what’s overkill. Real-world numbers would help a lot
[ Removed by Reddit on account of violating the content policy. ]
I tried Dokie AI again after seeing that it now supports Image 2, and honestly the output looks a lot better than I expected.
The biggest difference is the visuals. A lot of AI presentation tools can generate slides fast, but the images often feel like random placeholders. With this update, Dokie AI seems to create visuals that fit the slide topic much better.
I tested it with a business-style deck and a marketing presentation, and both looked more complete than before. The images were more relevant, the slides felt more consistent, and the whole deck looked closer to something I could actually edit and use.
It still needs some manual polishing, of course. I would still check the wording, add real data, and adjust some slides before sharing it. But as a first draft, it feels much stronger now.
For anyone who makes reports, pitch decks, marketing decks, or work presentations often, Dokie AI’s Image 2 update is worth checking out.
[ Removed by Reddit on account of violating the content policy. ]
[ Removed by Reddit on account of violating the content policy. ]
Ever show a model photo to your barber and it just… doesn’t turn out like you imagined? A picture might show the style you like, but it rarely tells them how to handle your face shape, hair density, crown, or fringe.
I’ve found that a really useful haircut brief should answer a few things before you sit down:
I’ve started using a tool called Aurcue to help me visualize hairstyle directions from my own photo. It’s not about letting AI decide my cut, it’s just a way to make sure I can explain clearly what I want and why to my barber.
Of course, the barber still has to consider cowlicks, growth patterns, and hair texture. But compared to walking in with a single Pinterest photo and hoping it translates, having a small visual guide really helps.
So, when you ask for haircut advice, what details do you wish people included besides just photos?
Sometimes I genuinely feel like Spotify understands my life better than people in real life do.
Not in a creepy way, but more like, “why is this Daylist exactly what I needed at 11 p.m. on a random Tuesday?” I open Spotify thinking I’m doing fine, and then it recommends some oddly specific playlist like soft nostalgic evening or anxious indie commute, and I’m like… okay, apparently that’s where I am emotionally right now.
Wrapped makes this feeling even stronger every year. It doesn’t just show me what songs I listened to. It feels more like a recap of what kind of year I survived. Some songs are tied to work stress, some to random phases of confidence, some to late-night overthinking, moving on from someone, or those few weeks when one album basically became my entire personality.
It records life less as entertainment and more as all those ordinary little moments. In a way, it tracks my emotional timeline better than I expected.
[ Removed by Reddit on account of violating the content policy. ]
[ Removed by Reddit on account of violating the content policy. ]
Today, we are thrilled to open-source Ling–2.6–1T from the Ling family.
Tailored for real–world, complex scenarios, this trillion–parameter model introduces targeted optimizations across inference efficiency, token overhead, and agentic capabilities, making it highly effective for coding and daily workflows.
Ring-2.6-1T is a 1T-parameter-scale thinking model with 63B active parameters, built for real-world agent workflows that require both strong capability and operational efficiency. It is optimized for coding agents, tool use, and long-horizon task execution, delivering leading results on benchmarks including PinchBench, ClawEval, TAU2-Bench, and GAIA2-search.
With adaptive reasoning effort across high and xhigh modes, Ring-2.6-1T dynamically allocates reasoning budget based on task complexity. This enables stronger performance with lower token overhead, especially in tool-heavy and multi-turn agent workflows.
Ring-2.6-1T is designed for advanced coding agents, complex reasoning pipelines, and large-scale autonomous systems where execution quality, latency, and cost efficiency all matter.
p1-2: Erhai Lake, Dali
p3: Yunshan Ping, Lijiang
p4: Blue Moon Valley, Lijiang
Just got back from a week in Yunnan with a friend. We did Kunming, Dali, Lijiang by train. Every stop was worth it.
Kunming was mostly food. Stone Forest and the Ethnic Village were solid, but I keep thinking about the wild mushroom hot pot. The server brought it over, set a timer on the pot, told us not to touch it. There's a mushroom in there called jianshouqing that's apparently toxic undercooked. So we sat there, hungry, watching soup. When the timer went off I grabbed a bowl of broth and it barely looked like anything special but the flavor was deep. I think this is what Chinese people call xian.
Dali was where we slowed down. Erhai Lake was stunning, wedding photoshoots everywhere, can't blame them.Second day I planned my route based on recommended spots from EasyGoChina and checked out Seven Dragon Maidens Pools and the Chongsheng Three Pagodas. Best part though was a bonfire after dinner. Ethnic minority aunties led everyone in traditional dances, tourists, locals, old, young, everyone holding hands around the fire. None of us knew the steps. Perfect. A local guy told me they have a Torch Festival in summer that goes way harder. Coming back for that.
Lijiang was the one that got me. Jade Dragon Snow Mountain from Spruce Meadow looked fake. Photos don't do it. Couldn't make it to the summit because altitude hit me hard, but I'd bought an oxygen canister ahead of time. No shame. Then Blue Moon Valley. The water was this shade of blue that looks edited in photos. It's not.
Spring weather was real, sunny every day, perfect temps. But wear sunscreen. The UV at altitude is no joke. Got home and my family said I was a full shade darker. If you're on the fence about Yunnan just go.
We are currently pushing a new feature for our SaaS app, but our marketing budget is basically non existent. Right now, the team is completely stuck in the classic Quality vs. Quantity debate for our video content.
If we go the Quality route: Is the ROI actually going to be there? We are terrified of sinking two or three days into polishing a single video, only for it to totally flop with the algorithm and bring in absolutely zero conversions.
If we pivot to Quantity: We get way more shots at hitting the right audience in the traffic pool. But how on earth do we maintain that kind of high frequency output? If we stick to traditional editing workflows, we will never hit our volume goals. Plus, we are super paranoid that spamming out low effort, mass produced garbage will completely tank our brand trust and credibility.
After debating in circles, our team is leaning towards the Quantity approach just to get data and test the waters. But honestly, we are pretty lost on how to actually execute the roadmap from here.
Has anyone here been in this exact spot? How do you actually scale your production to keep up with high volume targets? Are there any solid workflows or specific tools you guys are using to pump out product promos without losing your minds?
Would love to hear your insights!
Game Title:
*[planning to name it soon, suggestions welcome!]*
Playable Link:
https://luddi.ai/play/cmor3q8100001wr6nruyce5qd
Platform:
Web / Browser-based [Mobile, mostly]
Description:
This is a minimalist, fast-paced arcade game centered on a single core mechanic: survival. Players must control their avatar to successfully navigate through a continuously generated, randomized series of obstacles. The primary goal is to stay alive for as long as possible, but the game is designed to test your reflexes and pattern recognition to their limits. The difficulty curve is dynamic, meaning the game's speed, the density of the obstacles, and the complexity of the paths will steadily increase as you survive longer. It starts off deceptively simple, giving you time to learn the basic rhythm, but it quickly evolves into a high-pressure, skill-testing gauntlet. This isn't just about moving; it’s about anticipating, reacting instantly, and maintaining focus under duress. The constant acceleration creates a feeling of mounting tension where even the smallest miscalculation ends the run, driving a strong "one more try" loop.
Free to Play Status:
Involvement:
I am a member of the development team that created this game. My current role includes development and playtesting, but I'm currently stuck at the same point, so I'm hoping to see if others can beat my high score.