tar.gz

[Akshay 🚀 @akshay_pachaar]

LLMs vs. Jev, clearly explained!

TL;DR

The key difference is not that Jev generates faster.

Jev does not generate text at all.

A traditional LLM receives context and produces an answer one token at a time. Even when the output is a small JSON object, every token depends on those generated before it.

Jev receives the same context but evaluates predefined decisions directly. When those decisions are independent, it can evaluate all of them in parallel.

Consider an agent handling a failed deployment. It may need to determine:

→ Whether the incident is urgent
→ Which team should handle it
→ Whether the proposed command is risky
→ Whether the task is complete

An LLM generates a response containing these answers sequentially. The application then parses and validates it.

With Jev, you define the questions and expected answer types upfront. It evaluates them together and returns typed answers with probabilities.

Jev supports three decision primitives:

1. **Choice** selects from known options, such as engineering, billing, or sales.

2. **Score** places the input on an ordered scale, such as low, medium, or high risk.

3. **Noul** evaluates a yes-or-no condition and returns the probability that it is true.

The probabilities matter as much as the selected answers.

If engineering receives 91% probability and billing receives 9%, automatic routing may be reasonable. If the probabilities are 52% and 48%, the system can escalate, gather more context, or call a stronger model.

This keeps control inside ordinary software.

Code owns the thresholds and consequences. Jev supplies the semantic judgment that a normal `if` statement cannot derive from unstructured text.

It works best when the possible answers are known, the decision depends on meaning, and a careful person could judge the input quickly.

It is not designed for writing, summarization, code generation, arithmetic, or decisions requiring several dependent reasoning steps. Independent questions can run in parallel, but decisions that depend on earlier results must remain sequential.

Jev also cannot return an option outside the declared schema, but it can still select the wrong valid option. Type safety prevents malformed outputs, not incorrect judgments.

The clean mental model is this:

LLMs generate new language when the answer space is open.

Jev evaluates known paths when the answer space is bounded.

I wrote the full breakdown explaining Jev and where it fits.

The article is quoted below.

[tar.gz @tar47182990]

@threejs Awesome! Is this pure Three.js, or did you also use Blender for the 3D models?

[Token Gremlin @TokenGremlin]

GPT-6 Astra vs DeepSeek V4.1 Flash in Three.js.

Same kind of task, two completely different approaches. The visual difference alone is pretty interesting.

Source footage and original material: @OpenDesignHQ

[Graalito @Graalitoo]

#astra cooked up this hunting demo in savannah in the browser with #threejs

https://t.co/VseAL4hQ7o

[Bilal Khan @Bk23544]

After 200+ hours of work, Three.js Grassworks is finally ready to try.

The interactive demo is live:
https://t.co/MImqQODiPx

It is launching soon. Join the waitlist.

The first 50 to join the waitlist and buy Grassworks get the full demo code+assets for free.

#threejs #webgpu

[tar.gz @tar47182990]

@oguzthedev what engine do you use, bro

[zac.eth @zacxbt]

11 hours left in the @commonsmade grand prix 🏁

top 48 now racing: https://t.co/X8x7a9iaa7

[Oguz @oguzthedev]

I vibecoded a web browser game with Claude, it was played by 10k+ users less than a month!
Now working on a new version for Steam!
Play for free now! https://t.co/7HDgusRj2q
#beneloil

[Vibe Coding Thailand @vibecodingth]

Steve Yegge ทำโรงงานซอฟต์แวร์ชื่อ Wheelhouse ที่ใช้ AI agent หรือโปรแกรม AI ที่ทำงานเองได้ ราว 50 ตัวช่วยกันเขียนโค้ด

วันหนึ่งเขาเพิ่งรู้ว่า agent สร้างระบบกฎหมายขึ้นมาเอง ทั้งรัฐธรรมนูญและศาล รวม 450 ชิ้น โดยไม่มีใครสั่ง

ข้อสรุปของเขาสวนทางกับวิธีคุม AI ที่เราคุ้นเคย

1/11

[Chucky @chukinice]

This is my submission for this week's @spawn JAM, Ultimate Destruction

Built in just 3 days. Eat the city, chomp the enemies and defeat the boss

One multiplayer lobby, 3 single player cities, 3 different bosses

[AssetSaleLove @汗人柱 @AssetLove]

この連続斬り、最後の火柱まで気持ちよすぎる🔥
​
左右へ薙いで、回転斬りからド派手なフィニッシュ！
炎・水の斬撃やメテオなど、全12種のトゥーン戦闘VFX
​
『Epic Toon VFX 4 - Fire and Water effects』
$25.30 => $7.59 70%OFF⚡️8/23まで
https://t.co/miuVQoOfe6
#Unity ゲーム開発 VRChat ワールド制作

クーポン適用で、
$7.59 => $5.69 さらに25%OFF

🎟️ 初購入限定25%OFFクーポン
​​
クーポンコード1：PD89HNM81S
※最大 5回使える
​​
クーポンコード2：RTNEW60TNH7B
​​
🎟️ さらに25%OFFクーポン✨

クーポンコード1：NEW1QJXJ7B
​
クーポンコード2：ED89SYBU5X

👇【 11日目】本日のフラッシュセールはこちら(20件)
https://t.co/2yitRmOl2V

[Norvex @norvex1029]

This is f**king dangerous

A normal student just bought an iPad and Mac Mini for $2,200.

Connected them to his MacBook.

Three computers on one desk — his roommates thought he was mining crypto.

He just set up the automation and went to sleep.

By morning, the system had already processed hundreds of leads, written personalized emails, and filled the CRM without a single touch.

The team that used to handle the same workflow cost $7,000 a month.

He paid $2,200 once.

The crazy part isn't the hardware.

It's that three machines on one desk can now run an entire AI workflow while you're asleep.

No assistant.
No manual follow-up.
No sitting in front of the screen.
He just set it up and let the machines work.
Bookmark this before your feed buries it.

Read the full guide below.

[tar.gz @tar47182990]

@GameFxlab @threejs @webgl_webgpu 🤩how you build that model

[Bjarne Øverli @iamdothash]

A concept that's been brewing for a while. Cliamp is TUI first, and that will not change. What do you think of a native GUI wrapper? Ensuring low memory usage.

I love it myself.

[pixel_salvaje @Pixel_Salvaje]

PixZels 0.007 is out!

Here are some updates:
https://t.co/30klmmr3rv

#pixelart #Gamedev @itchio

[Maket @Maketplans]

Invisible wall toggle + auto-furnish button

Feels like magic 🪄