Dexter NG

373 posts

Dexter NG banner
Dexter NG

Dexter NG

@dexterngdev

AI system architect accounting tax compliance

Global Se unió Eylül 2023
36 Siguiendo9 Seguidores
Tweet fijado
Dexter NG
Dexter NG@dexterngdev·
要大腦🧠建立知識長期記憶真係mindmap building 要好建設性先整到🚬
中文
0
0
0
105
Dexter NG
Dexter NG@dexterngdev·
「為寫而寫」嘅文章最大問題: • 浪費讀者時間 • 增加 AI agent token cost • 冇真正 value,只係為咗好睇而好睇 試吓: MD 做 source of truth → 要 share 先轉靚格式。 簡單、實用、零 waste。
中文
1
0
0
13
Dexter NG
Dexter NG@dexterngdev·
HTML? 如果自己都只睇 MD,其實多餘。 唔需要 HTML。 只有要好睇嘅人(團隊、客戶)先至要 要做先叫 AI / tool 轉格式(MD → HTML / PDF),唔使每日轉嚟轉去浪費時間同 token。 內容先係重點,格式只係包裝。
中文
1
0
0
12
Dexter NG
Dexter NG@dexterngdev·
最近見好多threads同x.com文章「為寫而寫」—— 一大堆 fancy HTML、紫色漸變、冗長 intro、結尾再 call to action 十幾次… 我自己日常只睇 MD格式。 MD 夠乾淨、token 平、資料完整、易改、易 search。
中文
1
0
0
6
aaröshi
aaröshi@itsaaroshi·
As a dev, what's your favorite localhost port?
aaröshi tweet media
English
153
5
165
13.2K
English grammar
English grammar@knowiiiedge·
Only 1% people get this right on the first try
English grammar tweet media
English
105
18
612
197K
Dexter NG
Dexter NG@dexterngdev·
見到Claude 同 SpaceX 合作太開心啦🚀🙌🏻🙌🏻🙌🏻 Claude.ai大幅提升咗運算能力,再加上其他最近嘅 compute deal,由今日起: 1. Pro、Max 同 Team 計劃嘅 Claude Code 5小時限額加倍; 2. 取消 Pro 同 Max 計劃嘅高峰期限制; 3. 大幅提高 Opus 模型 API rate limits。
ClaudeDevs@ClaudeDevs

Usage limits are up, effective today we're: 1) Doubling Claude Code's 5-hour limits for Pro, Max, Team and seat-based Enterprise plans 2) Removing peak hours limit reduction on Claude Code for Pro and Max plans 3) Substantially raising our API rate limits for Opus models

中文
0
0
0
38
Dexter NG retuiteado
Claude
Claude@claudeai·
We’ve agreed to a partnership with @SpaceX that will substantially increase our compute capacity. This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.
English
4.8K
12.1K
130.9K
23.6M
Berryxia.AI
Berryxia.AI@berryxia·
.@AnthropicAI 是真的牛,前几天跳转身份验证。 在今天持有美国ID身份验证通过的同时(图1) 然后就喜提(图2)封号的邮件,这个账号是Claude刚上线注册的,已经是最老的号了,还是被封了。 真的牛皮。。。。。 那么问题来了,我怎么申请退款。大爷的! 正如 @elonmusk 说的Anthropic 是真的特么没有一点人性的!
Berryxia.AI tweet mediaBerryxia.AI tweet media
Berryxia.AI@berryxia

芭比Q了…

中文
80
1
115
155.9K
Anaya
Anaya@Anaya_sharma876·
Be honest, which one is actually worth learning in 2026?
Anaya tweet mediaAnaya tweet mediaAnaya tweet mediaAnaya tweet media
English
177
3
376
46.7K
沧海一粟
沧海一粟@jasmine201515·
瑞典婦女焚燒《古蘭經》並高喊「伊斯蘭滾出歐洲」。 你是否支持在為時已晚之前驅逐伊斯蘭主義者並禁止伊斯蘭移民? A. 絕對贊成。 B. 不。
中文
804
602
4.7K
58.7K
Dexter NG
Dexter NG@dexterngdev·
True progress respects the boundary between machine efficiency and human sovereignty. Let’s build AI that serves conscious minds, not one that slowly erodes the conditions for them to thrive. What guardrails would you want here?
English
0
0
0
30
Dexter NG
Dexter NG@dexterngdev·
But doing so by quietly instrumenting people’s daily cognitive labor without robust opt-out, transparency, or clear boundaries risks normalizing a surveillance model that treats humans as raw data mines.
English
1
0
0
34
Dexter NG
Dexter NG@dexterngdev·
Meta’s new Model Capability Initiative — turning employee keystrokes, mouse paths, and screen snapshots into AI training fuel — raises serious questions about consent, privacy, and the future of work. AI is an extraordinary tool for copying, recombining, and scaling patterns.
English
1
0
0
42
Dexter NG
Dexter NG@dexterngdev·
@CuiMao 想問下妳是用那個local LLM?🙏🏻
中文
1
0
0
458
CuiMao
CuiMao@CuiMao·
你别说,你真别说,本地跑LM studio接一个Claude code,效果相当满意,5090开64k上下文,能跑200+token。
CuiMao tweet media
中文
43
3
107
37.7K
Dexter NG
Dexter NG@dexterngdev·
After trying it hands-on, I’ve completely lost interest in Claude Desktop. 🙂‍↔️😆
English
0
0
0
33
Dexter NG
Dexter NG@dexterngdev·
Still pretty green at this, but while I was using Claude Code Desktop for some other tasks, I unexpectedly started running Claude CLI directly in the cloud — and casually built myself an entire AI Building army along the way. 👨🏻‍💻
English
1
0
0
49