Z

1.2K posts

Z banner
Z

Z

@edgerunner_z

Swing trader | Amateur quant trader Don't follow my trades or you may be rugged. Always respect the market. Always DYOR

Katılım Eylül 2022
1.5K Takip Edilen190 Takipçiler
Z
Z@edgerunner_z·
@leyten maybe we can put the current inference to openrouter first? @leyten
English
0
0
0
43
leyten
leyten@leyten·
Demo for the sharded inference engine on: shard.c0mpute.ai gpt-oss-120b on 4090's, max input context 95k, max output context infinite (tok/s degrades with context) queue of max 12 prompts concurrently, use it wisely
English
70
81
361
80.2K
Pump.fun
Pump.fun@Pumpfun·
vibe shift.
Pump.fun tweet media
English
697
242
1.9K
159.6K
Z
Z@edgerunner_z·
@blknoiz06 solana:EmcxFTNVDqyLHp11NvwvLZ4D7LKGbG9i7B8RF7dwpump and solana:jtojtomepa8beP8AuQc6eXt5FriJwfFMwQx2v2f9mCL
Filipino
0
0
0
9
Ansem
Ansem@blknoiz06·
can only hold 2 solana eco coins & 2 solana memes for the next 6 months alongside $SOL which do you pick and why
English
2.2K
163
2.7K
549.6K
Z
Z@edgerunner_z·
@noobnooc 去年聖誕節買，360$一年max，不知道今天聖誕節還有沒有
中文
0
0
0
1.9K
Nooc
Nooc@noobnooc·
智谱也不便宜啊，是我进的方式有问题吗
Nooc tweet media
中文
115
0
112
75.3K
Z
Z@edgerunner_z·
@0xSero seems there is no way to opt out from training?
English
1
0
1
619
0xSero
0xSero@0xSero·
Just ported al my work over to zCode.
0xSero tweet media
English
36
1
217
18.7K
Z
Z@edgerunner_z·
@andrewmccalip the extension keeps crashing
English
0
0
0
96
Z
Z@edgerunner_z·
@Debuggerdam seems like it keep crashing when using multiple instances? and very not stable even for only one instance
English
1
0
0
21
Debuggerdam
Debuggerdam@Debuggerdam·
so a little update. i've earned almost 9$ in 24 hours on kickbacks.ai if it can keep this up, that would mean an extra (free) income of almost 270$ per month. meaning atleast 1x claude code pro and 1x codex pro accounts for free. amazing deal if you ask me.
Debuggerdam tweet media
Debuggerdam@Debuggerdam

This shit is just funny and very exciting at the same time. You literally earn money while waiting on GPT 5.5 to finish his tasks. First time ever I've wanted Codex to be slow lol I've just installed this and got it working on Windows 11 in Cursor with Codex CLI.

English
1
0
1
616
Z
Z@edgerunner_z·
@Stv_Lynn Premium+ 自帶的也可以用composer嗎？
日本語
1
0
0
2K
Steven Lynn
Steven Lynn@Stv_Lynn·
SuperGrok 近期有优惠，30刀/3个月，两天后截止 可以买来蹭 composer2.5
Steven Lynn tweet media
中文
43
4
76
37.2K
Fantasy only
Fantasy only@Fxx44564·
@andrewmccalip got my stripe wired up! nothing through yet though email is rhodriellis99@gmail.com sry to nag lol, great work btw. No rush ofc
Fantasy only tweet media
English
2
0
4
948
BlockedPath
BlockedPath@BlockedPaths·
Kimi K2.7 did not impress me in the slightest. No where near 5.5 or deepseek.
English
27
0
54
14.7K
Z
Z@edgerunner_z·
@buildwithhassan Should be weighted by monthly limit of the models
English
0
0
2
321
Hassan
Hassan@buildwithhassan·
opencode published their real model usage data. what developers actually run when they're paying for it: 1. deepseek v4 flash: 32T tokens 2. deepseek v4 pro: 19T tokens 3. kimi k2.6: 6.5T tokens deepseek is running more tokens than the next 16 models combined. it's actual usage from developers spending their own money. glm-5.1 grew 419% too. the models winning on price and reliability aren't always the ones winning on twitter.
Hassan tweet media
English
29
25
505
34.3K
Z
Z@edgerunner_z·
@buildwithhassan glm-5.2 seems to be on the next level
English
0
0
0
243
Hassan
Hassan@buildwithhassan·
my personal chinese coding model tier list from daily usage: tier 1: GLM-5 ≈ GLM-5.1 ≈ Kimi K2.6 > Qwen 3.7 Max (somehow the most expensive and not the best) tier 2: Qwen 3.7 Plus (under 200K context) > DeepSeek V4 Pro ≈ V4 Flash ≈ MiMo 2.5 Pro not there yet: MiniMax, Hunyuan testing K2.7 Code soon and GLM-5.2. ranking might change by morning.
English
39
6
420
36.4K
Andrew McCalip
Andrew McCalip@andrewmccalip·
Get paid to wait The Claude Code spinner might be the most watched line on Earth. So I turned it into an ad marketplace. Advertisers bid on it. You keep 50% of the money. Install the extension → get cash from ads. Introducing Kickbacks
English
1.1K
500
13.2K
7.7M
Ivan Fioravanti ᯅ
Ivan Fioravanti ᯅ@ivanfioravanti·
Grok Build + MTPLX v1 + Qwen 3.6 27B Optimized Quality🚀 Did you know you can use Grok Build with local models? I did not before this experiment 🤷🏻‍♂️ Here it is in a video!
English
4
6
70
8.6K
Z
Z@edgerunner_z·
@danielhanchen is it possible to run it as an api endpoint?
English
1
0
1
188
Daniel Han
Daniel Han@danielhanchen·
We made DiffusionGemma run via llama.cpp locally! It works well with Unsloth GGUFs and you can run it in realtime visualization mode or normal chat CLI mode! See our docs unsloth.ai/docs/models/di… on how to set it up!
Unsloth AI@UnslothAI

Google releases DiffusionGemma.✨ The new 26B-A4B diffusion text model runs locally on 18GB RAM. It supports high-speed text generation, thinking, image, video and 256K context. Run and train via Unsloth Studio. GGUF: huggingface.co/unsloth/diffus… Guide: unsloth.ai/docs/models/di…

English
11
38
397
35.2K
Z
Z@edgerunner_z·
@0xSero But it is using 4x the memory of qwen3.6
English
0
0
2
955
0xSero
0xSero@0xSero·
We're all sleeping on Step-3.7-Flash. It's phenomenal.
0xSero tweet media
English
51
19
578
41K
Z
Z@edgerunner_z·
@bourneliu66 codex spark 好像速度也是1000TPS
中文
0
0
0
565
刘小排
刘小排@bourneliu66·
MiMo-V2.5-Pro UltraSpeed的速度快得有点太过分了，代码写的太快就像龙卷风…… 和其他厂商不一样，小米这是Pro模型直接出来1000token/s，不是flash模型。 恐怕在世界范围内也是第一次吧。从来没见过这么快的速度。 网页版在这里platform.xiaomimimo.com/ultraspeed 有API
中文
166
26
411
87.1K
Z
Z@edgerunner_z·
@AdamShao 一直通過不了gpt security的KYC
日本語
0
0
0
3.8K
Adam Shao
Adam Shao@AdamShao·
让 AI 自己写了一个自动挖 bug 的矿机，token 就是电力，bug 就是区块奖励。每个月挖矿成本 200 美金，每天早上起来收菜。不幸的是，crypto 项目方自己都没钱了……
Adam Shao tweet mediaAdam Shao tweet media
中文
39
2
78
77.1K

Keşfet

@leyten @Pumpfun @HopiumPapi @blknoiz06 @c0mputeAI @noobnooc @0xSero @andrewmccalip