[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@TheAhmadOsman Avatar @TheAhmadOsman Ahmad

Ahmad posts on X about ai, llms, build a, gpu the most. They currently have XXXXXX followers and XX posts still getting attention that total XXXXX engagements in the last XX hours.

Engagements: XXXXX #

Engagements Line Chart

Mentions: XX #

Mentions Line Chart

Followers: XXXXXX #

Followers Line Chart

CreatorRank: XXXXXXX #

CreatorRank Line Chart

Social Influence

Social category influence technology brands XXXXX% stocks XXXX% finance XXXX% celebrities XXXX% social networks XXXX% currencies XXXX%

Social topic influence ai 13.41%, llms #56, build a #1165, gpu #525, ollama #42, open ai #847, inference #92, agentic #239, llamacpp #12, strong XXXX%

Top accounts mentioned or mentioned by @franzferdinan57 @gerry @jasonth0 @gradient_hq @grok @itsjustmarky @meowmoyo @testtm7873 @artale93 @theonekev @annanidev @mjs527 @mxmzb @wgreklek @test_tm7873 @zlorthishen @theodorganescu @dropustoday @maziyarpanahi @brooksy4503

Top assets mentioned Cloudflare, Inc. (NET) Wink (WINK) Alphabet Inc Class A (GOOGL)

Top Social Posts

Top posts by engagements in the last XX hours

"nobody should use ollama btw slower than llama.cpp on windows slower than mlx on mac slop useless wrapper alternatives lmstudio llama.cpp exllamav2/v3 vllm sglang parallax (esp. for pooling devices together) like literally anythings better than ollama lmao"
X Link 2025-11-23T15:20Z 32.1K followers, 94.7K engagements

"people that think like this still exist btw and theyre in for a big surprise sadly"
X Link 2025-11-25T20:40Z 32.1K followers, 558.9K engagements

"TEN HOURS LEFT TO WIN DGX Sparks & Mac Minis ur submission can be as simple as sending hello world msg while showing ur laptop or as complicated as a room full of hardware u have to use ur office chair as a laptop stand reply with ur submissions so i dont miss them"
X Link 2025-12-07T19:37Z 32.1K followers, 26.1K engagements

"the RTX PRO 6000 is becoming my absolute favorite card God if only i could afford putting together a 14x RTX PRO 6000 server would be a feat to behold"
X Link 2025-12-09T06:21Z 32.1K followers, 5888 engagements

"Chinese AI Labs Tier List: S: ByteDance Deepseek Moonshot Zhipu (Zai) A: MiniMax Qwen B: Baidu StepFun Tencent C: InternLM Meituan RedNote Xiaomi Western labs tier list will be posted once Prime Intellect and Arcee AI drop their models"
X Link 2025-11-23T16:19Z 32.1K followers, 33.8K engagements

"just a reminder: RAM and storage didnt suddenly get expensive because of AI supplies are being artificially capped on purpose this is textbook price-gouging nothing more"
X Link 2025-11-24T18:19Z 32.1K followers, 20.5K engagements

"extremely bearish signal wining & dining potential hires happens either out of generosity or desperation in this case its the latter Mark Zuckerberg will be remembered for hiring Alexandr Wang making him boss of Yann LeCun & destroying FAIR peaked w/ Llama3 ig"
X Link 2025-12-02T19:41Z 32.1K followers, 13.6K engagements

"youre OpenAI hire a small army of ex-Meta ad and monetization people a Slack channel just for ex-Facebook staff brings in the full targeted ads playbook launch a browser users install it and OpenAI collects personalized granular data at scale its a browser-shaped surveillance device its a mapping machine of your workflows its a reverse-engineering tool for the internets data pipelines deployed at scale via their users launch Sora X a TikTokstyle social network infinite AI-generated video feed you create or remix clips upload your face become the cameo star every scroll like remix is another"
X Link 2025-12-02T23:17Z 32.1K followers, 19.8K engagements

"i used to drop all my LLM/AI educational threads right here on the timeline but to keep everything organized (and actually findable) im moving all of it past and future to Substack if youve been learning from my posts consider subscribing"
X Link 2025-12-05T20:30Z 32.1K followers, 1791 engagements

"you waifu wouldn't disappear like this every time cloudflare goes down if she's hosted locally anon don't let her go Buy a GPU"
X Link 2025-12-05T09:26Z 32.1K followers, 11.1K engagements

"local llms XXX running a model = inference (using model weights) inference = predicting the next token based on your input plus all tokens generated so far together these make up the "sequence" tokens words they're the chunks representing the text a model sees they are represented by integers (token IDs) in the model "tokenizer" = the algorithm that splits text into tokens common types: BPE (byte pair encoding) SentencePiece token examples: "hello" = X token or maybe X or X tokens "internationalization" = XX tokens context window = max tokens model can "see" at once (2K 8K 32K+) longer"
X Link 2025-12-05T18:05Z 32.1K followers, 98.8K engagements

"the Buy a GPU movement gets a little stronger every single day 💪"
X Link 2025-12-05T23:03Z 32.1K followers, 2966 engagements

"so every piece of slop useless software is getting donated to the Linux Foundation now or what first Googles Agent2Agent now Anthropics MCP about time though thank God were done with this trash protocol"
X Link 2025-12-10T06:44Z 32.1K followers, 11.3K engagements

"some of anthropic rugpulls so far: X years retention of all conversations and code all data will be used for training 1.58-bit quantized models during daytime plus not getting opus X in claude code max plans limits cut in half X weeks ago no comms weekly limits without concrete numbers 5x/20x plans being actually 3x/8x of plus DMCA takedowns of repos that have to do with Claude Code one of which was my own personally windsurf no access to claude X cutting off openai api access what an absolutely horrible company"
X Link 2025-08-29T07:14Z 32.1K followers, 464.7K engagements

"MASSIVE llama cpp now now ships with a built-in web UI stop using ollama there are no more excuses"
X Link 2025-11-04T16:51Z 32.1K followers, 96.2K engagements

"do you know that you can win a DGX Spark or a Mac Mini this holiday season anon (literally just a photo gets you in) - run a model on your laptop link up all your home pcs or build a remote cluster with friends pcs & laptops - build an ai agent + video demo it"
X Link 2025-11-20T22:19Z 32.1K followers, 46.4K engagements

"be you want to actually learn how LLMs work sick of just start with linear algebra and come back in X years decide to build my own roadmap no fluff. no detours. no 200-hour generic ML playlists just the stuff that actually gets you from whats a token to I trained a mini-GPT with LoRA adapters and FlashAttention goal: build fine-tune and ship LLMs not vibe with them. not "learn the theory" forever build them you will: build an autograd engine from scratch write a mini-GPT from scratch implement LoRA and fine-tune a model on real data hate CUDA at least once cry keep going X phases if you"
X Link 2025-11-21T14:38Z 32.1K followers, 90.2K engagements

"ollama max plan $100/month XXX premium requests premium models include Gemini X Pro yes they have people paying $XXX for XXX premium requests to generally available models lol the grifting is strong with the ollama folks"
X Link 2025-11-22T19:45Z 32.1K followers, 83.6K engagements

"you can see it running on device XX vram peaks around 21.9gb at the end of each smoothing pass then drops to 5gb when the next layer starts host ram swings between 90-150gb per layer iteration should be done in 6hrs give or take"
X Link 2025-11-28T04:06Z 32.1K followers, 6494 engagements

"1 of the worlds only X DRAM vendors is gone Samsung and Hynix pulling back from consumer supply will do data centers only this will impact the entire storage stack DRAM supply tightening SSD performance & pricing get hit NAND demand shifts HDD market follows"
X Link 2025-12-03T21:39Z 32.1K followers, 16.2K engagements

"- you are - a random CS grad with X clue how LLMs work - get tired of people gatekeeping with big words and tiny GPUs - decide to go full monk mode - X years later i can explain attention mechanisms at parties and ruin them - heres the forbidden knowledge map - top to bottom how LLMs actually work - start at the beginning - text tokens - tokens embeddings - you are now a floating point number in 4D space - vibe accordingly - positional embeddings: - absolute: i am position X - rotary (RoPE): i am a sine wave - alibi: i scale attention by distance like a hater - attention is all you need -"
X Link 2025-10-04T11:00Z 32.1K followers, 245.3K engagements

"be anthropic engineer realize long-running agents still have goldfish memory every new context window = new intern who forgot everything from yesterday project goes from build a clone of ChatGPT to why is half the frontend missing again agents try to one-shot entire apps run out of context mid-feature next session wakes up like boss who touched the router folder why is the server on fire other times claude walks in sees X buttons rendered declares the whole project complete packs up its laptop goes home humans dont work like this engineers leave breadcrumbs notes git commits tests todo lists"
X Link 2025-11-26T18:59Z 32.1K followers, 545.3K engagements

"this is what Anthropic wanted that AI cyberattack article was the setup they walk into congress with it push their anti-opensource agenda into regulation make it harder for people to own their intelligence resist by making sure everyone knows NOT to trust them"
X Link 2025-11-27T00:05Z 32.1K followers, 61.9K engagements

"top X local LLMs to run at home for December 2025 MiniMax-M2 at #1 great for agentic workflows thanks to interleaved thinking very solid at coding & good at UI/design as well fits on 2x RTX PRO 6000 or 8x RTX 3090s w/ full context (specifically MiniMax-M2-AWQ-4bit) GLM-4.5-Air at #2 all-around MVP & daily driver also when running multiple agents in parallel this is my go-to uses fewer GPUs & less memory for KVCache which allows me to spin up more agents fits on a single RTX PRO 6000 or 4x RTX 3090s w/ full context nearly as good as big brother however not as power hungry (specifically"
X Link 2025-12-05T23:46Z 32.1K followers, 19K engagements

"be cloudflare bouncer of the entire internet guarding XX% of all HTTP requests on the planet wake up on december X just gonna bump this buffer from 128kb to 1mb what could possibly go wrong 500s detonate across the planet XX% of the internet immediately starts screaming internet: bro AGAIN cloudflare sweating reverting pretending everything is normal be engineers trying to fix a React RSC vuln accidentally unlock a secret global-outage speedrun FL1 proxy hits a bug thats been chilling undetected since Obama era lua throws an exception so hard it becomes a lore event killswitch meant to save"
X Link 2025-12-06T03:34Z 32.1K followers, 184.7K engagements

"you are a person neck-deep in LLMs wading through a sea of jargon shards experts layers across nodes tensor parallelism pipeline parallelism data parallelism you keep hearing epic claims like X million tokens/sec on 1024 GPUs and wonder what the hell does that even look like on real hardware we'll start with what layers even are then walk the five parallelism axes and then show how they combine is every GPU running a full copy of the model or are the weights sharded across the cluster are layers sliced up what are layers to begin with a layer = attention + MLP + residual glue a transformer"
X Link 2025-12-07T23:52Z 32.1K followers, 9646 engagements

"its called local inference T. just quantize a 13B toss it on your 3090 and let that thing cook context windows are for people who rent compute"
X Link 2025-12-09T05:14Z 32.1K followers, 7086 engagements

"@brooksy4503 Bash is all you need Agents already know how to use Bash tell them to write code execute and run commands"
X Link 2025-12-10T08:09Z 32.1K followers, XXX engagements

"all Super Flower PSUs are on sale this Cyber Monday (plus a gift card & cooling fans) on Newegg if youre planning an AI build nows a good time to lock one in im just waiting on RAM + CPU to get delivered then its straight into recording some awesome build videos"
X Link 2025-12-01T20:43Z 32.1K followers, 3755 engagements

"this is nothing theyre just starting ill keep saying it: Buy a GPU and learn to use LLMs locally"
X Link 2025-12-08T19:07Z 32.1K followers, 12.1K engagements

"everythings starting to come together this case is one of several youll see in the upcoming Buy a GPU build-guide videos im insanely hyped for it (its about to become my main driver too) the videos were about to shoot are going to be absolute fire"
X Link 2025-12-03T02:29Z 32.1K followers, 4612 engagements

"closing the night with this honestly the best feeling the whole point of Buy a GPU was to get people to stop panicking and actually learn the stack every time i get a message like this it reminds me its worth it even with all the noise and attacks good night everyone :)"
X Link 2025-12-03T05:23Z 32.1K followers, 6256 engagements

"one more RTX PRO 6000 Buy a GPU keeps on winning 💪🔥"
X Link 2025-12-04T00:17Z 32.1K followers, 3553 engagements

"tone look im only tellin you this cuz people been talkin' so you hear it from me first yeah i bought a gpu alright a big one the kind you need two hands for XX gigs things basically a space heater they had it in stock for once what you want me to do walk away from that you dont ignore signs like that but i swear on my muddas tit i wasnt buildin a data center or working on skynet behind your back i am runnin no operation i just wanted some clean local inference maybe run a model that doesnt put emdashes everywhere and do a lil fine-tune for the kids i was home by XX anyway after what happened"
X Link 2025-12-04T03:26Z 32.1K followers, 2364 engagements

"THIS THING IS MASSIVE everybody say hello to the case for my new build (EPYC 9004 CPU w/ 512GB DDR5 build + a yet-to-be-announced number of RTX PRO 6000) gonna be sharing more pictures as things get delivered i am so hyped for this build and the Buy a GPU build(s) video guides"
X Link 2025-12-04T19:30Z 32.1K followers, 13.3K engagements

"@prv_khvr skip ollama"
X Link 2025-12-05T03:20Z 32.1K followers, 2846 engagements

"asuka & i are giving away DGX Sparks & Mac minis this holiday season how to win one easy build a local AI lab with Parallax build a cool agent or app with local LLMs using Parallax submit a video of either (or both) p.s. asuka isn't competing with this project;)"
X Link 2025-11-30T00:42Z 32.1K followers, 26.7K engagements

"DANGEROUS STUFF over XX million ChatGPT conversations are now evidence. a judge just ordered OpenAI to hand them over in a lawsuit if anything ever screamed Buy a GPU run your AI locally protect your damn data its this right here"
X Link 2025-12-04T18:16Z 32.1K followers, 658.5K engagements

"you are a random CS grad with X clue how LLMs work get tired of people gatekeeping with big words and tiny GPUs decide to go full monk mode X years later i can explain attention mechanisms at parties and ruin them heres the forbidden knowledge map top to bottom how LLMs actually work start at the beginning text tokens tokens embeddings you are now a floating point number in 4D space vibe accordingly positional embeddings: absolute: i am position X rotary (RoPE): i am a sine wave alibi: i scale attention by distance like a hater attention is all you need self-attention: who am i allowed to pay"
X Link 2025-12-04T23:48Z 32.1K followers, 80.2K engagements

"my top X local LLMs to run at home for December 2025 MiniMax-M2 at #1 great for agentic workflows thanks to interleaved thinking very solid at coding & good at UI/design as well fits on 2x RTX PRO 6000 or 8x RTX 3090s w/ full context (specifically MiniMax-M2-AWQ-4bit) GLM-4.5-Air at #2 all-around MVP & daily driver also when running multiple agents in parallel this is my go-to uses fewer GPUs & less memory for KVCache which allows me to spin up more agents fits on a single RTX PRO 6000 or 4x RTX 3090s w/ full context nearly as good as big brother however not as power hungry (specifically"
X Link 2025-12-05T04:56Z 32.1K followers, 25.4K engagements

"The Ahmad Lab will be a frontier lab"
X Link 2025-12-09T19:05Z 32.1K followers, 3517 engagements

"ngl i miss ChatGPT XXX Pro now that i downgraded to plus membership didnt realize how much i was offloading to it until it was no longer an option"
X Link 2025-12-10T19:30Z 32.1K followers, 4427 engagements