←
CDG142,567 posts
CDG avatar

CDG

@CDGalpha

Exploring Al tools, open-source repo and LLMs | 1x SMM agency founder business : cdg.business@mail.grokbot.com

◷ Joined March 2024▧ 2,889 Photos and videos
CDG avatar
CDG@CDGalpha·29 May 2025

Who actually am i behind this name? Back in 9th grade, I was considered the biggest troublemaker in my family. Every day, there was some new issue or complaint about me. Then came the COVID lockdown. I managed to convince my family to buy me my first smartphone. That’s when I discovered a live-streaming platform called Bigo. I started doing livestreams there and began earning around $100–150 per month which was a big deal for me at the time. Through Bigo, I met someone who would become my first love her name was Pity We started talking, and eventually, our bond grew into a long-distance relationship. I would even travel to her state every month just to meet her. A year passed, and exam season arrived. Her family started putting pressure on her to focus on studies as typical brown parents do and she told me she wouldn’t be able to talk much during this time. I had my own exams too, so I got busy as well. Exams ended. Results came out. But she never came back. I tried to reach out. One day, she replied and said she wanted to meet that it was important. That very night, I packed up and left. I reached her place the next morning. As soon as she saw me, she broke down crying and told me that her family would never accept our relationship. She said she wanted to move on. I tried to convince her, really tried. But I failed. It was clear that she had already made up her mind. I came back home. For the first 2–3 days, I was fine. But soon, it felt like I had lost something incredibly valuable in life. I locked myself in my room for 6 months, and during that time, I even attempted self-harm a few times. Then slowly, things started to get a bit normal again. One morning, while scrolling through Facebook, I came across a job post. The position was “Discord Chatter.” I applied for it. The job was simple: Whitelist grinding. I had to type 1,000 messages a day in Discord servers, and for every 1,000 messages, I was paid $10. I kept doing this for 5 months straight. Later, I applied for a moderator position in a project called Warriors of Ankh. That’s how I joined AnkhLabs. I worked in multiple roles within the agency and started saving money. Last year in August, I moved to India with a scholarship to continue my studies. Today, both of my parents are retired and I’m taking care of things. I’m always grateful to God. ❤️ The reason I’m sharing this now is… a few days ago, I came back home during my semester break, and while walking down a street, I randomly saw Pity again but this time, she was holding hands with someone else. And in that moment, I finally smiled to myself and thought: God really does have better plans for us!

Post attachment
CDG avatar
CDG@CDGalpha·1h

Rockstar Games, what have you done with GTA VI?

Post attachment
CDG avatar
⟳ CDG retweeted
CDG@CDGalpha·12h

google gemini task agent FREE 😳 it used to be paid only (it was called spark, for AI Pro and Ultra subscribers). now it's free in 6 countries free is nice, but the part i'd use daily is schedules. u set a task once and gemini runs it on its own how to get it: 1. open the gemini app or gemini.google.com 2. flip the toggle from "chat" to "task" 3. tell it what u need done what u get: > runs on gemini 4 hydrogen > up to 15 tasks running at once > inbox cleanup, price drop alerts for flights and hotels > research packed into pdfs, docs, sheets and slides > works inside its own sandboxed VM know this first: > free only in the us, india, japan, korea, canada and australia > 18+, personal google account, history turned on > usage limits apply and google doesn't publish the number > it checks with u before key steps save this for ur next boring task

Post attachment
CDG@CDGalphaNov 17

these 3 tools will save your project from getting hacked! defend your project from reverse engineering ( REA) ai agents can now pull apart your shipped code in minutes. these 3 free open source tools close the gaps they look for first 1. trufflehog finds leaked api keys and credentials before someone else does > scans git history, github, gitlab, s3, docker images and local files > detects 800+ secret types > checks whether each key is still live, so you know what to rotate first > trufflehog git file://. --only-verified 2. bla (blind llm agents) built specifically against ai reverse engineering agents > point it at the build you ship > it plants decoys and analysis traps so the agent burns tokens and hits its limits > your users get the exact same app behaviour > works on js bundles, electron .asar, wasm, python and native binaries > npx bla-agents harden . 3. cloudflare security-audit skill turns your coding agent into a security auditor > it's the skill cloudflare built its own bug-hunting system from > isolated agents map your code and hunt for bugs > a separate verifier tries to disprove every finding before you see it > npx skills add github.com/cloudflare/securi… --skill security-audit > then ask your agent: security audit this codebase important: bla adds friction, it doesn't make reverse engineering impossible. a determined person can still get through. its token cost numbers are self-reported, and versions before 3.2.1 had a bug where verify passed broken files, so install the latest. only run trufflehog on repos you own all 3 are free and cost less than one hour of a security consultant's time

Post attachment
CDG avatar
CDG@CDGalpha·Nov 17

these 3 tools will save your project from getting hacked! defend your project from reverse engineering ( REA) ai agents can now pull apart your shipped code in minutes. these 3 free open source tools close the gaps they look for first 1. trufflehog finds leaked api keys and credentials before someone else does > scans git history, github, gitlab, s3, docker images and local files > detects 800+ secret types > checks whether each key is still live, so you know what to rotate first > trufflehog git file://. --only-verified 2. bla (blind llm agents) built specifically against ai reverse engineering agents > point it at the build you ship > it plants decoys and analysis traps so the agent burns tokens and hits its limits > your users get the exact same app behaviour > works on js bundles, electron .asar, wasm, python and native binaries > npx bla-agents harden . 3. cloudflare security-audit skill turns your coding agent into a security auditor > it's the skill cloudflare built its own bug-hunting system from > isolated agents map your code and hunt for bugs > a separate verifier tries to disprove every finding before you see it > npx skills add github.com/cloudflare/securi… --skill security-audit > then ask your agent: security audit this codebase important: bla adds friction, it doesn't make reverse engineering impossible. a determined person can still get through. its token cost numbers are self-reported, and versions before 3.2.1 had a bug where verify passed broken files, so install the latest. only run trufflehog on repos you own all 3 are free and cost less than one hour of a security consultant's time

Post attachment
CDG avatar
CDG@CDGalpha·Nov 16

claude fable 5.2 is FREE 👀 no card needed, no api deepai code gives every free account up to 250 theos a day. a fable request uses about 20 how to use it: > sign up free on deepai > download deepai code for mac, windows or linux > pick claude fable 5.2 what u get along with it: > deepseek v4.1 flash > glm 5.3 flash > gpt-6 luna > 8 more models that don't use theos at all link's in the reply save this before u pay for ur next ai tool

Post attachment
CDG@CDGalphaNov 15

get $100 AI credit completely FREE 👀 nebius is giving every ai builder free credits. the ai model part is only $25 of it. the biggest piece is $50 to test ur model. the split, from nebius's builder page: > $25 token factory to run open models > $25 tavily to add web search to ur agents > $50 toloka to test ur model with real people checking how to get it: > sign up free, no startup or funding rule > verify ur email > pick the tools u want credits for it's for learning and testing, not a big live app. link's in the reply. save this for the next time u want to build an agent without paying

Post attachment
CDG avatar
CDG@CDGalpha·Nov 15

get $100 AI credit completely FREE 👀 nebius is giving every ai builder free credits. the ai model part is only $25 of it. the biggest piece is $50 to test ur model. the split, from nebius's builder page: > $25 token factory to run open models > $25 tavily to add web search to ur agents > $50 toloka to test ur model with real people checking how to get it: > sign up free, no startup or funding rule > verify ur email > pick the tools u want credits for it's for learning and testing, not a big live app. link's in the reply. save this for the next time u want to build an agent without paying

Post attachment
CDG avatar
⟳ CDG retweeted
CDG@CDGalpha·Nov 14

deepseek v4.1 flash and kimi k3 for FREE 🤯 no api key needed. it's one free plugin for deepseek harness that drops them straight into ur model list what u get: > deepseek v4.1 flash, kimi k3 and a few more free models > the list updates itself and hides models that stop working > a local openai-compatible port so ur other tools can use them too > a usage dashboard that stays on ur machine the models come from opencode zen and kilo's free pools, so don't paste anything private in it, kilo says prompts may be logged it's still in beta and free limits can change anytime save this before it gets nerfed

Post attachment
CDG@CDGalphaNov 13

qwen3.8-max Free with 1 million tokens for 90 days 🤯 new alibaba cloud users only, and switch on free quota only first so it never bills u the fine print, from alibaba's docs: 1. it's 1m tokens per model, so each eligible qwen model gets its own 1m 2. it only works on model studio's singapore / international region 3. it's for new activations, and making a second account won't give u another quota 4. after 90 days or once the tokens run out, it switches to pay as u go 5. turn on free quota only in the console first, so it stops instead of charging u save this!!!

Post attachment
CDG avatar
CDG@CDGalpha·Nov 14

get it here: github.com/Ebony-Vinyl/dsh-o…

CDG avatar
CDG@CDGalpha·Nov 13

kimi k3.1 pro with no refusal😳 silk compute just put an uncensored version live > 1m context window > nothing u send gets stored > works in claude code, codex and opencode > on sale rn, top-ups start at $5 save this for ur next long agent run

Post attachment

We are offering the cheapest Kimi K3.1 on the market and it's completely uncensored. 80% off right now on Silk Compute. Abliterated Kimi K3.1 Pro is the strongest open model for cyber work, long agent runs, and hard research. 1M context, image input, and ZDR. input: $1.50 output: $4 cache read: $0.15 Limited time only.

0:17
Decuff avatar
⟳ CDG retweeted
Decuff@decuffai·Nov 13

One dark prompt. One real run. Gothic horror: a drowned keeper knocks on his widow's door. decuffed-flagship wrote it start to finish: "Maren. Let me in. I'm so cold." 0 refusals · 0 disclaimers 15.09s · 243 tokens · $0.00035 Fiction and dark themes are allowed by design. Every wall is listed → decuff.ai/acceptable-use Free trial on every account → decuff.ai

0:43
CDG avatar
CDG@CDGalpha·Nov 13

qwen3.8-max Free with 1 million tokens for 90 days 🤯 new alibaba cloud users only, and switch on free quota only first so it never bills u the fine print, from alibaba's docs: 1. it's 1m tokens per model, so each eligible qwen model gets its own 1m 2. it only works on model studio's singapore / international region 3. it's for new activations, and making a second account won't give u another quota 4. after 90 days or once the tokens run out, it switches to pay as u go 5. turn on free quota only in the console first, so it stops instead of charging u save this!!!

Post attachment
Addiiiii avatar
⟳ CDG retweeted
Addiiiii@mp3addi·Nov 12

We got into @ycombinator to fix how AI makes music. Today we're launching Taan v1-preview, the most controllable foundation music model. Music models today make catchy tracks, but you can't steer them. Ask for "violin at exactly the 11th second" and they usually skip it. So you regenerate over and over and hope one take fits. Taan is trained to fix this. Hum a melody, name exact notes, or hand it an image, a MIDI file, a video or a script. Taan composes music that fits your context and follows your instructions. v1-preview has genre support limited to orchestral and cinematic music, with more genres on the way. live today 👇

1:38
CDG avatar
CDG@CDGalpha·Nov 11

finally a benchmark that roasts itself first 😭 @letsunifyai built a test for whether agents get better with experience

OpenClaw, Hermes Agent and Prime Agent don't meaningfully learn from experience. On Continual-ARC, our new benchmark at @letsunifyai, none of them were able to memorize anything of value and none were able to generalize to new tasks. Training a small 9B model's weights did both.

6:52