Sabitlenmiş Tweet
アイゼン様
3.9K posts

アイゼン様
@NipponJackson
AI開発者、AIプロジェクトマネージャー - 現在、アニメの嫁さんを募集中です。DMを送るか、私の投稿に返信してください。
Earthling Katılım Ocak 2024
1.4K Takip Edilen38 Takipçiler

@weswinder Gemini models are highly censored, and mostly useless, unless you are using it to summarize things and use it as small RAG replacement. Even the Google Devs use Claude code or Codex heavily. So again, what's the point of Antigravity ide at all. LoL!
English

@weswinder since 2 months Google team is trying to force their Antigravity ide users to use Gemini models, but nobody likes those models. From what I came to know, it's a discounted tool to get access to claude models for Indians.
English

@shekhar_tw Nobody uses gemini cli. They should have focussed fully on AG 2.0 instead of highly censored omni and other bogus tools. AG 2.0 is lacking on many things. @kevinhou22 like wtf ! Just copy claude code and cowork.
English

What Google actually launched yesterday:
> Google killed Gemini CLI and replaced it with bloated Antigravity agents
> Gemini 3.5 Flash benchmarks look fake, real-world coding is disappointing
> Gemini 3.5 Flash pricing is insanely high - $1.50/$9 per million tokens
> Google turning everything into expensive half-baked agents
kill useful tools, hype shiny new stuff, charge more.
English

@TheWhizzAI I went to India a few months ago, and now for werid reasons my account is getting tagged as Indian, with Indian level nonsensical censoring, even the bikini mention in prompt triggers censoring of images LoL! . Wth is going on there.
English

@NipponJackson over-censoring is Google's biggest self-inflicted wound in AI video
English

Google spent billions of dollars on Gemini Omni.
It still can't make a person do a backflip.
I tested it
→ Asked Gemini to generate a backflip
→ Physics was broken
→ Body looked wrong
→ Tried 3 times. Same result.
A billion-dollar model.
One simple move. Still failing.
Don't trust the hype. Test everything yourself.
Bookmark this.
Comment on your worst AI fail.
English

@OfficialLoganK @antigravity Everyone knows even their own Deveopers, managers, PMs and all use claude code instead of AG. I wonder why !! GTFOH!
English

Very excited for all the stuff the @antigravity team has been cooking :)
English

6 Reasons Why Fans Are Disappointed With The Boys
- Homelander doesn’t feel scary anymore
- Too much plot armor for The Boys
- The focus shifted from Supes vs Boys to random side conflicts
- Starlight and Hughie Campbell lost their chemistry
- Repetitive powers and boring fight scenes
- Feels more like setup for spin-offs than a proper ending


English

Easy prediction for The Boys final episode:
- Starlight will kill The Deep at the beginning of the episode.
- Marie tries to take out V1 from Homelander’s body, but Homelander kills her.
- Kimiko tries to blast Homelander, but he escapes.
- Soldier Boy teams up with The Boys
- Soldier Boy takes away Homelander’s powers, and Butcher kills him.
- Later, Kimiko blasts Homelander again, and he dies slowly.
- Butcher turns evil and wants to kill all the supes, he doesn’t want another Homelander in the future. - He kills Soldier Boy.
- Hughie fights Butcher, and Ryan kills Butcher




English

@swapnakpanda The AI actually is much powerful but as per experts since few weeks with proofs, the closed AI companies have been nerfing almost all models in the name of security and indirect lawsuits. So in a way founder was right but he cannot prove it in long term. Just enjoy Gemma 4B.
English

A founder in India fired an employee in just 4 weeks after hiring.
The founder used Lovable and accused the employee isn't performing at even 50% of his speed.
The same Lovable that was found yesterday of a big cyber security breach.
Who will tell these founders, AI may run faster, but you need a human to build a secured production-grade system?



English

@HealthRanger Sonnet and opus 4.6 both are unusable since almost 2 weeks now.
English

I'M CALLING THIS OUT: There is a concerted, top-down effort to dumb down all AI models that are accessible by the public, including open source AI and cloud-based AI. And there's a reason why this is happening. It explains why Opus 4.7 is such a disappointment, among other things. The chasm between in-house frontier AI / AGI versus open-to-the-public AI is widening. On purpose. I'll have a full report on this tomorrow or the next day.
English

拿网上已经泄露的 claude design 系统提示词,
然后把这个提示词的链接丢给 claude code,同时输入:
---
请根据这个提示词,帮我生成一个设计 skill,并且命名为 /sydney-design
---
然后就可以用 /sydney-design 在 claude code 等 coding agent 里面用 skill 方式调用 claude design 了。
同时你可以宣称你已经逆向了 claude design,并且“开源”来引流。
应该还可以再发一篇公众号。




Sac@Saccc_c
昨天爆火的Claude Design的系统提示词已经泄漏了。 看了一眼,提示词里重视整体设计context、鼓励多方案探索、预先设置好了去AI味的规则。 有点牛逼,完整版放下面了👇
中文

Chris Camillo reveals how people are making $500K/year being an "AI guy" for small businesses
"There are millions of small businesses out there and almost none of them are willing to embrace AI right now, you just walk in and say give me one area where you're leaking money, I'll fix it for free"
"Within days you've set up an AI agent answering their calls, sending responses, getting quotes out in real time… and you just increased their revenue by 5 to 15% at essentially no cost"
"Now you're invaluable. They're paying you $2,000 or $3,000 a month to be their AI guy, you replicate that across 10 or 20 businesses and you're making half a million a year"
English
アイゼン様 retweetledi

I put the entire Claude Code GTM Engineering Playbook into ONE Notion doc.
8 sections. No fluff.
- How to get set up correctly from day one: Pro plan, terminal install across Mac, Linux, and Windows, GUI install via Antigravity or VS Code, and bypass permissions mode
- What to put in your project brain file, what to leave out, and how to get Claude to update it automatically when it keeps making the same mistake
- How to run plan mode step by step and when to skip it for simple tasks
- How to build a skill file from scratch, fix one that keeps failing, and install 5 GTM skills worth building first: lead scraping, email labeling, proposal generation, outbound sequence writing, and client onboarding
- MCP install process, token cost checks after every install, the best MCPs for GTM work, and how to cut token usage by 50 to 100x by converting MCPs into skills
- Sub-agents and agent teams: the 3 cases where they earn their cost, reliability math for parallel runs, and how to enable parallel variant exploration
- What is eating your context before you type anything, how to use /compact and /clear correctly, and model selection for parent vs sub-agents
- Modal deployment: any skill as a live URL in under 2 minutes, form interface setup, and connection to n8n, Make, or Zapier
This is the setup I would have KILLED for before spending months piecing together how to actually get productive in Claude Code from documentation, YouTube tutorials, and scattered GitHub threads.
Like + comment "CODE" and I'll send it over
(must be connected for priority access)

English

@shio_no_kokyu あなたのプロフィールを見ると、あなたは定期的に男性と性行為をしているようですが、なぜ彼らと結婚するのが難しいのですか?あなたは結婚相手として、ベータタイプの男性を探していますか?
日本語






