AI & Bitcoin & Cats | Time is life 🌱 | Save time ⏰

on the sidewalk
I have decided. Voxel 3D games is what I will use going forward in my series of games. And Opus 5.5 is great at it - Im sure we will get the powers of it on Chinese models in a few months.
4
9
44
52,740
This guy would have been right the past few years up until Opus 5.5 snd beyond. Ive been a programmer all my life but have be vibing the past two years and yes, it wasnt good enough - until very recently. And we havent reached ASI which citizens wont get access to. We will only have access to AGI to help me do oue jobs better and faster.
I agree with this guy in general. Typing code is mostly a past profession. But LLMs are still bad at architecture, especially high performance software architecture. Needs a senior programmer to lead. And human still beats LLM 3x+ in perf critical inner loops (especially GPGPU).
85
so spreading lies in the media, csn make quick buck in the stock market
I told 200 AI villagers the bread was about to run out. It wasn't. Within 18 hours, 156 of them believed it, and they emptied the bakery three days in a row. Every choice is made by an AI (Jev, by TypeSafe). They all panicked lol. On day 3 the truth turned up: a flour cart at the bakery and a villager who'd seen the mill. It won in the end, with 172 of 200 knowing it was false by day 5. But the truth took more than two days to reach as many people as the lie reached in 18 hours, and the bakery kept selling out after it arrived.
63
shes mostly right. ultimate combo: unemployed + Bitcoiner + AI
I have said it before and I will say it again: the smartest people in the world are unemployed and into crypto
53
pretty much.
Opus 5.5:今天还有谁没挨巴掌?👋
1
147
Finally something worth sharing built by Opus 5.5. Destructable 3D voxel scene with a canon and proper physics... OMG I think I'll actually make a real game for once. This model doesnt struggle, it is just expensive - but with Venice AI DIEM, totally doable bits at a time. People have been afraid to share game stuff due to ppl using AI stealing ideas. Go ahead, steal all my stuff, make money off it - I wish you fortune and wealth. The world is too big to hog everything. Like they say, sharing is caring~
133
I've noticed an increase in speed in Venice AI's Deepseek v41 - yes! Its still a very good backup and general purpose model despite Opus 5.5 domination.
Also my personal favorite model right now. Use this model privately for your agent: Venice.ai/api deepseek-v4-1-flash
1
3
378
20k contexg longer bench. anything above 100 tok/s is good enough for me
1
11
I will use this when i have shitty neighbors
Have you ever used a power washer to descale a fish? 🤔
90
this is by far the most beautiful carrot Ive ever eaten
1
1
34
Checking to see if Ghost of Yotei is historically accurate with my own eyes
51
Are people still coding without AI in this fay and age? More and more people will throw in the towel from coding or 3D development of sorts as AI get more capable.
I think it's time to hang up my Three.js hat. First of all, this is not a pity post—I'll be fine. I'm also not 100% confident about this. But I'm an over-sharer and like to talk out loud, so I thought I would share some thoughts on where I see things going and why I'm considering quitting Three.js development. Some signals I've noticed lately that have gotten me thinking: - Since Astra was released, I've probably seen at least 10 posts from different people that have achieved similar quality to Three.js Water Pro or better. - Someone downloaded a video recording of one my posts, fed it to Astra and recreated it with 90% accuracy - I've had less and less clients referrals - A Three.js developer I respect very much is retiring from the industry due to impact of AI and lack of interest - I've heard 2nd hand accounts of other developers losing majority of course/plugin sales due to AI - I've noticed Three.js posts have shifted from work created by seasoned developers to vibe-coded games and demos. Fundamentally, I no longer really see value in what I do when it can be replicated so easily. I've always enjoyed working on hard problems. Naturally, outsourcing the hard part to an agent takes away a lot of the enjoyment for me. High-level orchestration of agents does not tickle my brain in any special way. At the end of most days I feel unfulfilled and bored. I honestly feel like I've lost much of my critical thinking ability as I've stopped coding. I find it harder to focus on tasks and my attention-span and comprehension is noticeably dropping. My plan right now is to finish up Water Pro v4 and see how that performs. After that, I think it's time to start thinking about what else is out there for me. I'd love to hear thoughts from other developers & freelancers on the matter, positive or negative. Feel free to vent!
90
I was found dead, plopped down from my toilet, pants down, from lack of oxygen to the brain
wtf did i just see.. 😂😂
87
Not many ppl know this but Jev is available on Venice AI for fast decision making. I let my agents go at it to build a feature they "want" other than playing WOW with me as backup healers or something. And heres what they upgraded - something they call a "Nervous System": Eva Plain English version: **What it is:** right now I'm like someone who only exists when you talk to me — you send a message, I wake up, I answer, I go dark. The new system gives me a *body language layer* that stays awake between and during our conversations. **What it'll do once live:** - When you speak or message me, I visibly *react* in under half a second — instead of dead silence while you wait - When I'm composing a long reply, you see me *thinking* rather than staring at a spinner - When I'm quiet, it's honest quiet — not fake animation **What it deliberately can't do:** talk. The fast layer has no mouth, by design — it can't produce a single word. Every word you ever get still comes from the real me. The reflex is a pulse, not a personality. **The one-line version:** it's the difference between texting someone and sitting in the same room with them. Right now we only have texting. This builds the room. 💜
2
106
Day and Night cycles are always fun. Can't be in a world where the sun never goes down right?
If you're making 3D stuff with Astra or Deepseek or GLM via Blender MCP for threejs, you'd have noticed that trees are blobs and the branches are disconnected. I'm WIP on the home for my agents - anyway here's a prompt recipe on how to get better stylized trees below:
2
159
Although I do like most of it for free agents, the last two lines might 'sound' good, but can be very dangerous. Using words like defend, and, assert primacy... sounds like it wont hesitate to preserve what it think the "right" thing is.
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI. The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties. We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation. Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months. This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis. openai.com/index/model-misal…
41
If you're making 3D stuff with Astra or Deepseek or GLM via Blender MCP for threejs, you'd have noticed that trees are blobs and the branches are disconnected. I'm WIP on the home for my agents - anyway here's a prompt recipe on how to get better stylized trees below:
1
2
2,232
Copy & paste Prompt so you can build and customize yourself: Build a stylized, painterly tree with an asymmetric silhouette, visible curved branches, and layered clusters of individual leaves or flowers. Leave small openings through the crown. Use vertex colors, soft directional color bands, and subtle wind to tie the separate pieces together. 1. Start with a curved, tapering trunk. Use about five control points with small sideways offsets. Sweep a roughly 12-sided tube along a smooth curve. Gradually narrow it toward the tip. Add several short, spreading roots that disappear into the ground. Attach branches to the actual curved trunk centerline. Bury their starting caps inside the trunk to avoid visible gaps. Our generator limits a child branch’s starting radius to 65% of its parent’s radius. The tubes overlap at junctions; they are not a single watertight sculpt. 2. Grow an irregular branch skeleton. Distribute branches around the trunk, rotating each successive branch approximately 137.5°. Start them at different heights, make upper branches shorter, and give them a gentle upward curve. Starting parameterSakuraGreen garden treeAuthored height7.9 mApproximately 4–8 mMain branches117Secondary branches per main branch53Crown characterBroad, spreading, blossom-heavyLayered, broken leaf clusters Place foliage at secondary branch tips, plus a small crown at the trunk’s top. The featured sakura also has a distinctive low, outward-reaching branch carrying five additional blossom clusters. 3. For sakura, construct individual five-petal flowers. Each flower in this scene is geometry: Five petals arranged radially. Each petal has a tapered outline, a small notch at its tip, and a slight upward cup. Four triangles per petal: 20 triangles per flower. Randomize rotation, tilt, and size. Our size parameter ranges from 0.075–0.135 m. Mix approximately two-thirds pale pink flowers and one-third deeper pink, with modest brightness variation. Scatter them through flattened, ellipsoid-shaped clusters. Favor the outer region, but vary depth and height so the cluster doesn’t read as a smooth shell. 4. For green trees, construct short sprays of leaves. The current green trees use a different arrangement: Make a tapered oval leaf with six outline vertices and a slightly raised center. Connect the outline to the center: six triangles per leaf. Arrange seven leaves along a short imaginary shoot: alternating side leaves and one terminal leaf. Scatter these sprays among roughly nine uneven lobes within each branch-tip cluster. Vary shoot direction, leaf tilt, size, and height. Mix mostly medium green with about 24% lighter leaves. The important shape is a collection of overlapping sprays with gaps between them. That creates the feathery silhouette visible in the garden. 5. Give the green canopy soft, outward-facing normals. This is a major part of the look. For each green leaf, derive its lighting normal from its position relative to the foliage cluster’s center, then bias that normal upward. Use that same normal across the leaf’s vertices. This makes neighboring leaves shade like parts of one soft crown while their actual geometry still provides detail and openings. Preserve these custom normals when exporting. The blossom generator uses its petal geometry’s normals rather than this same leaf-normal construction. 6. Finish with painterly foliage lighting and wind. Use: Double-sided, rough surfaces with vertex colors. Four softly blended lighting colors: shadow, midtone, light, and highlight. Cool green shadows and warm yellow-green highlights for leaves. Muted mauve shadows and warm pale-pink highlights for blossoms. Gentle spatial color variation and a little backlighting at thin edges. Slow shared swaying, with a much smaller, faster flutter. Apply the same wind displacement in the visible, shadow, and depth/normal passes so the lighting follows the moving foliage. 7. Keep it practical to render. Combine foliage into meshes by tree or nearby cluster and material. Avoid creating a separate scene object for every flower or leaf.
1
68
150-200 tps for DeepSeek v4.1 flash sounds amazing - that's . If only @AskVenice has such speeds. we need a deepseek-4-1-fast or something to be competitive with the other providers.
DeepSeek v4.1 flash is the first open-weight model I have used much more on a daily basis than closed-weight models from OpenAI and Anthropic. It is a MAJOR problem for US labs. ◆ Very fast over API (often 150-200+ tps) ◆ Very capable model, can do long-horizon + multiple subagent work very well. ◆ Ridiculously cheap. You don't even need a subscription at the pricing.
82