Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

A 27 YEAR OLD BANGALORE DEV BOUGHT A DECOMMISSIONED BANK SERVER RACK FOR $3,200 AT AUCTION AND NOW PULLS $24,000 A MONTH FINE TUNING LLMs FOR US SAAS COMPANIES raj is 27, bangalore, one bedroom flat with a ceiling fan, won the lot off a state bank IT refresh... in march, four 2U storage shelves with 96TB of enterprise SAS drives and a managed switch added two custom towers with RTX 3090s under the desk, runs llama 3.3 70B and fine tunes indic and english variants for 11 US saas clients on hugging face pause at 0:30 on the finger pointing at the yellow drive caddies, that is gear that sold new for $180,000 $24,000 in, $61 in karnataka power, the auction lot paid itself back in 4 days, no claude bill no rate limits, nothing leaves the flat the window is open, follow and bookmark before it closesshow more

Flandermaxx

1,301 subscribers

1,191,544 views • 3 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

33 Comments

Samar3 months ago

Utter BS !

Alex Rubenstein3 months ago

I don't think that decommissioned bank server will be able to handle this many requests, considering the server doesn't usually have GPUs.

Calder3 months ago

a bank server rack in a one bedroom flat means the kid is sleeping next to industrial cooling fans, that is the actual story

Mayur Panghaal3 months ago

Claim 4: Fine-tuning LLMs for 11 US SaaS companies from a flat Possible. Many AI consultants operate solo. However, "11 US SaaS clients" sounds like a marketing number. Questions you'd want answered: What kind of fine-tuning?LoRA? QLoRA? Full fine-tuning? How large are datasets? How often are jobs run? Are these actual companies or one-off Upwork gigs? The claim is not impossible, but lacks evidence. Credibility: Medium.

Uves Arshad (montr.io)3 months ago

Time to surf bank auction sites 🌊

Aryan Singh3 months ago

You mean this guy and not some random 27 yr old Raj. Get your facts correct please

Mayur Panghaal3 months ago

A 70B model is large. Two RTX 3090s provide: 24 GB VRAM each 48 GB total VRAM A quantized 70B model can run across multiple GPUs and system RAM, but performance depends heavily on: quantization level inference engine context size throughput requirements For experimentation, yes. For serving 11 paying US SaaS clients simultaneously, that's much harder. The story omits critical details: What quantization? What tokens/sec? What uptime? What SLA?

Anish Bhanushali3 months ago

Bescom will render this hardware useless with voltage fluctuations and powercuts before it can process a single token

Animesh Singh3 months ago

Abhi ruko zara, sitaraman ji khud aayengi tax lene.

Drake3 months ago

wait, can you actually buy a decommissioned bank rack at auction in india, is that public information?

Liquidden3 months ago

anyone on the timeline actually running an old bank rack at home, what is the noise level?

Mayur Panghaal3 months ago

@grok are these claims credible ? Can two custom towers with RTX 3090s under the desk, runs llama 3.3 70B smoothly ?

Ajay Prabhu3 months ago

@MoShahx07 Let me sound like an uncle and say : seekho kuch 😬😬

Ai LLM Peasant3 months ago

When they want to sell the old crap but still want to increase the demand in order to sell at highest prices. Enter the fake stories and fuuusss hype !

Leet3 months ago

idk why these retards compare llama 3.1 70B to claude models, your shitty local model is not even close to any claude model not even haiku

KarmaWarrior3 months ago

BS pro max 😅

Rakita 🧂3 months ago

It’s the crypto mining from your room hype all over again. This will definitely kill AI for sure

sanjay patel3 months ago

Can anyone explain in simple words what is this ?

Esha Gavaskar3 months ago

Any proof that he makes that much.. or just storytelling

Hana Trần3 months ago

whoa, 24k a month just fine tuning llms? 😮

Ash_kay3 months ago

i dont that is true. electricity prices in India are quite high

Thanveer Karim3 months ago

Over engineered piece of garbage , nowadays a single dgx spark or a mac studio can do much more

Paarivaala3 months ago

I have run qwen 3.6 28b mixture of experts model 128k context on a laptop RTX 4060 8gb with turboquant and well optimised using llama.cpp with 33 t/s , it works well but everything goes to trash when it’s about agentic coding. You’ll again go back to cloud based frontier models.

rohithje.near3 months ago

@NammaBESCOM doing this without commercial Meter look at this scam

Dinesh Ram 🇮🇳3 months ago

Even AI makes better engagement clips, than this fake bullshit

Md. Asaad Sayed 🪽3 months ago

Homelab goals 🛐

Tsera3 months ago

Rate limits are dead

Kong Trading 🦍3 months ago

Decommissioned bank servers running LLMs is the best hustle

ArŰSH3 months ago

People are really today just dumb or what?

Chetan Kale3 months ago

insane

Firas D3 months ago

Who is paying for fine tuned versions of llama 3.3 70b? Come on man

Kushagra Keshari3 months ago

There is a voltage stabilizer but no power UPS or BESS? 😂

Ritik Verma3 months ago

"The server rack is the easy part. The monthly invoices are the hard part."

Related Videos

A 24-year-old in his garage built a 4x RTX 4090 home lab and now runs 70B models at 35 tokens/sec for $0.18 per hour. No cloud bills. No rate limits. Full privacy. He spent $18,400 on GPUs, power, and a used server chassis. Power draw hits 1.8kW at load. His electric bill jumped $340 last month. The wins: Local Llama 3.1 70B with full context. Private RAG over his own docs. Fine-tuning runs overnight. Zero vendor lock-in. The pain: Noise like a jet engine. Heat turns the garage into a sauna. One PSU failure and the whole rig goes down. Scaling beyond 4 cards means new electrical work and serious cooling. He still runs it every day. The models answer faster than any API, and nothing leaves his network. Four glowing 4090s in the dark. The fan roar never stops. The power meter spins. Worth it.

A 24-year-old in his garage built a 4x RTX 4090 home lab and now runs 70B models at 35 tokens/sec for $0.18 per hour. No cloud bills. No rate limits. Full privacy. He spent $18,400 on GPUs, power, and a used server chassis. Power draw hits 1.8kW at load. His electric bill jumped $340 last month. The wins: Local Llama 3.1 70B with full context. Private RAG over his own docs. Fine-tuning runs overnight. Zero vendor lock-in. The pain: Noise like a jet engine. Heat turns the garage into a sauna. One PSU failure and the whole rig goes down. Scaling beyond 4 cards means new electrical work and serious cooling. He still runs it every day. The models answer faster than any API, and nothing leaves his network. Four glowing 4090s in the dark. The fan roar never stops. The power meter spins. Worth it.

Mirage

350,017 views • 2 months ago

A 24 BAY DELL POWEREDGE T550 TOWER SERVER FITS IN A HOME OFFICE WITHOUT A RACK, 384TB OF RAW SAS STORAGE HOSTS EVERY OPEN SOURCE LLM EVER RELEASED LOCALLY, THE HOME AI LIBRARY TIER UNDER YOUR MAP 00:25 the reseller spins the chassis around, "it's a 24 bay Dell T550 tower server man, look it's got a boss card for dual NVMe boot SSDs" a refurbished T550 ships at $900-1,100 with single Xeon Silver, 64GB DDR4 ECC and 24 hot swap SAS bays in a tower form factor that fits next to a desk without a server rack 24x 16TB SAS drives at $80 each on eBay equals 384TB raw storage for $1,920, the same array on enterprise Mac Studio Pro tier would cost $35,000 in thunderbolt JBOD enclosures Llama 3.3 70B is 42GB, Qwen3-235B is 110GB, DeepSeek-V3 is 100GB, Mistral Large is 80GB, the entire ollama public model library fits inside 4TB, you can mirror every open weights release for the next decade your map covers compute boxes that swap one model at a time, the T550 is the model warehouse tier that feeds them, a homelab operator wires the tower over 10GbE to a mac mini and the mac never waits 30 seconds to pull a swap the window is open, follow and bookmark before it closes

A 24 BAY DELL POWEREDGE T550 TOWER SERVER FITS IN A HOME OFFICE WITHOUT A RACK, 384TB OF RAW SAS STORAGE HOSTS EVERY OPEN SOURCE LLM EVER RELEASED LOCALLY, THE HOME AI LIBRARY TIER UNDER YOUR MAP 00:25 the reseller spins the chassis around, "it's a 24 bay Dell T550 tower server man, look it's got a boss card for dual NVMe boot SSDs" a refurbished T550 ships at $900-1,100 with single Xeon Silver, 64GB DDR4 ECC and 24 hot swap SAS bays in a tower form factor that fits next to a desk without a server rack 24x 16TB SAS drives at $80 each on eBay equals 384TB raw storage for $1,920, the same array on enterprise Mac Studio Pro tier would cost $35,000 in thunderbolt JBOD enclosures Llama 3.3 70B is 42GB, Qwen3-235B is 110GB, DeepSeek-V3 is 100GB, Mistral Large is 80GB, the entire ollama public model library fits inside 4TB, you can mirror every open weights release for the next decade your map covers compute boxes that swap one model at a time, the T550 is the model warehouse tier that feeds them, a homelab operator wires the tower over 10GbE to a mac mini and the mac never waits 30 seconds to pull a swap the window is open, follow and bookmark before it closes

Antid

18,526 views • 3 months ago

$THIS IS WHERE "THE CLOUD" ACTUALLY LIVES - A ROOM OF DELL STORAGE ARRAYS YOU RENT A SLIVER OF EVERY MONTH. that clip is a real datacenter aisle, rack after rack of dell powerstore arrays, every hexagon panel packed with drives. this is "the cloud." not a metaphor, this room, plus the power, cooling and staff keeping it alive. when you pay for cloud storage, you rent a few gigabytes of these arrays. forever. the bill grows with every photo. here's the part they'd rather you not connect. these enterprise drives run on a schedule. data centers swap them out long before they fail, by the pallet. those decommissioned drives get tested and resold as recertified, the same 20tb enterprise disk for a fraction of new. so the exact hardware humming in this room ends up in a nas on your desk, minus the monthly rent. no cloud subscription, no data on someone else's array, no bill that grows with your library. the uncomfortable part isn't that the cloud is huge. it's that "the cloud" was always just this room, and you can buy a piece of it once instead of renting it forever. follow RetroChainer for what the cloud doesn't want you to picture, and bookmark this before your next storage renewal.$

THIS IS WHERE "THE CLOUD" ACTUALLY LIVES - A ROOM OF DELL STORAGE ARRAYS YOU RENT A SLIVER OF EVERY MONTH. that clip is a real datacenter aisle, rack after rack of dell powerstore arrays, every hexagon panel packed with drives. this is "the cloud." not a metaphor, this room, plus the power, cooling and staff keeping it alive. when you pay for cloud storage, you rent a few gigabytes of these arrays. forever. the bill grows with every photo. here's the part they'd rather you not connect. these enterprise drives run on a schedule. data centers swap them out long before they fail, by the pallet. those decommissioned drives get tested and resold as recertified, the same 20tb enterprise disk for a fraction of new. so the exact hardware humming in this room ends up in a nas on your desk, minus the monthly rent. no cloud subscription, no data on someone else's array, no bill that grows with your library. the uncomfortable part isn't that the cloud is huge. it's that "the cloud" was always just this room, and you can buy a piece of it once instead of renting it forever. follow RetroChainer for what the cloud doesn't want you to picture, and bookmark this before your next storage renewal.

RetroChainer

34,236 views • 1 month ago

#PrincessDiana’s Revenge Dress is going up for auction at Sotheby’s-here’s a closer look at it. Diana wore this dress only one time, 29 June 1984 to the Serpentine Gallery in Kensington Gardens. It was the same night that Charles’ interview with Jonathan Dimbleby aired where Charles confessed to adultery. The dress, designed by Christina Stambolian, is a black crepe silk with ruching and a chiffon train. Diana had purchased it in 1991 for £900 (£2,149 in 2026) but hadn’t worn it because she felt it was “too daring”. That went out the window on that night 🤣 It was sold at auction before Diana’s death in June 1997 at the Christie’s auction in NYC where she sold 79 of her dresses for charity at #PrinceWilliam’s suggestion. The dress sold for £39,098 (£93,360 in 2026) to a Scottish couple. The dress goes back up for auction in Dec.

#PrincessDiana’s Revenge Dress is going up for auction at Sotheby’s-here’s a closer look at it. Diana wore this dress only one time, 29 June 1984 to the Serpentine Gallery in Kensington Gardens. It was the same night that Charles’ interview with Jonathan Dimbleby aired where Charles confessed to adultery. The dress, designed by Christina Stambolian, is a black crepe silk with ruching and a chiffon train. Diana had purchased it in 1991 for £900 (£2,149 in 2026) but hadn’t worn it because she felt it was “too daring”. That went out the window on that night 🤣 It was sold at auction before Diana’s death in June 1997 at the Christie’s auction in NYC where she sold 79 of her dresses for charity at #PrinceWilliam’s suggestion. The dress sold for £39,098 (£93,360 in 2026) to a Scottish couple. The dress goes back up for auction in Dec.

Princess CarParkle 👑

25,933 views • 10 days ago

THIS WHOLE 3-NODE CLUSTER RUNS PROXMOX, CEPH AND KUBERNETES - ON 10 WATTS TOTAL. that clip is three zima boards wired to one little switch, and the power meter reads 10 watts for the entire thing. on those three tiny fanless boards: proxmox running a real cluster, three nodes as one datacenter ceph spreading storage across all three, so a node can die and your data doesn't kubernetes on top, kube1/2/3, plus docker and high-availability failover this is the exact stack that runs in datacenters, shrunk onto a shelf. kill one node and the vms migrate to the other two. it heals itself. and it pulls less power than a light bulb. no server rack, no cooling, no power bill you'd ever notice. the uncomfortable part isn't that it's small. it's that "enterprise infrastructure" was mostly the rack and the noise, and none of that was the actual capability. follow RetroChainer for more ways enterprise gear shrinks to a shelf, and bookmark this before you rent another vps.

THIS WHOLE 3-NODE CLUSTER RUNS PROXMOX, CEPH AND KUBERNETES - ON 10 WATTS TOTAL. that clip is three zima boards wired to one little switch, and the power meter reads 10 watts for the entire thing. on those three tiny fanless boards: proxmox running a real cluster, three nodes as one datacenter ceph spreading storage across all three, so a node can die and your data doesn't kubernetes on top, kube1/2/3, plus docker and high-availability failover this is the exact stack that runs in datacenters, shrunk onto a shelf. kill one node and the vms migrate to the other two. it heals itself. and it pulls less power than a light bulb. no server rack, no cooling, no power bill you'd ever notice. the uncomfortable part isn't that it's small. it's that "enterprise infrastructure" was mostly the rack and the noise, and none of that was the actual capability. follow RetroChainer for more ways enterprise gear shrinks to a shelf, and bookmark this before you rent another vps.

RetroChainer

42,348 views • 1 month ago

NVIDIA'S BIGGEST COMPETITOR AMD BUILT A BOX THAT RUNS A 235B MODEL AND KILLS $200 OPENAI AND $200 CLAUDE CODE FOR $9 A MONTH amd's ryzen ai max+ 395 is the first x86 chip that runs a 200 billion parameter model on a single piece of silicon with 128gb of unified memory the gmktec evo-x2 runs qwen3 235b fully, deepseek v3 comfortably and llama 3.3 70b with headroom. on linux you get 110gb of usable vram out of 128gb amd claimed the chip beat an nvidia rtx 5080 by more than 3x on deepseek r1 inference. a lunchbox sized pc outrunning a $1,000 discrete gpu on a real ai workload a heavy ai user pays $200 for claude code max, $200 for chatgpt pro, $20 for cursor and $20 for gemini. that's $5,280 a year and the box pays itself off in 9 to 10 months ollama installs in one command and claude code points at localhost so nothing leaves the machine and nothing costs per request bookmark this and read the article below

NVIDIA'S BIGGEST COMPETITOR AMD BUILT A BOX THAT RUNS A 235B MODEL AND KILLS $200 OPENAI AND $200 CLAUDE CODE FOR $9 A MONTH amd's ryzen ai max+ 395 is the first x86 chip that runs a 200 billion parameter model on a single piece of silicon with 128gb of unified memory the gmktec evo-x2 runs qwen3 235b fully, deepseek v3 comfortably and llama 3.3 70b with headroom. on linux you get 110gb of usable vram out of 128gb amd claimed the chip beat an nvidia rtx 5080 by more than 3x on deepseek r1 inference. a lunchbox sized pc outrunning a $1,000 discrete gpu on a real ai workload a heavy ai user pays $200 for claude code max, $200 for chatgpt pro, $20 for cursor and $20 for gemini. that's $5,280 a year and the box pays itself off in 9 to 10 months ollama installs in one command and claude code points at localhost so nothing leaves the machine and nothing costs per request bookmark this and read the article below

starmex

93,580 views • 3 months ago

AMD'S NEW MINI PC RUNS CLAUDE CODE OFFLINE FOR $0/MONTH AND REPLACES $200 OPENAI + $200 CLAUDE CODE SUBSCRIPTIONS amd's ryzen ai max+ 395 is the first x86 chip that runs a 200 billion parameter model on a single piece of silicon with 128gb of unified memory the gmktec evo-x2 runs qwen3 coder 30b at 40 to 50 tokens per second and llama 3.3 70b at 20 to 25 tokens per second locally. on linux you get 110gb of usable vram out of 128gb setup takes 90 minutes. wipe windows for linux, install ollama with one command, pull the model, point claude code at localhost. same interface, same agent loop, nothing leaves your network a heavy ai user pays $200 for claude code max, $200 for chatgpt pro and $20 for cursor. that's $5,040 a year and the box pays itself off in 5 months after break-even every month is $420 that used to leave your account. three year math is roughly $13,000 not paid to ai companies plus zero rate limits on overnight agents bookmark this and read the article below

AMD'S NEW MINI PC RUNS CLAUDE CODE OFFLINE FOR $0/MONTH AND REPLACES $200 OPENAI + $200 CLAUDE CODE SUBSCRIPTIONS amd's ryzen ai max+ 395 is the first x86 chip that runs a 200 billion parameter model on a single piece of silicon with 128gb of unified memory the gmktec evo-x2 runs qwen3 coder 30b at 40 to 50 tokens per second and llama 3.3 70b at 20 to 25 tokens per second locally. on linux you get 110gb of usable vram out of 128gb setup takes 90 minutes. wipe windows for linux, install ollama with one command, pull the model, point claude code at localhost. same interface, same agent loop, nothing leaves your network a heavy ai user pays $200 for claude code max, $200 for chatgpt pro and $20 for cursor. that's $5,040 a year and the box pays itself off in 5 months after break-even every month is $420 that used to leave your account. three year math is roughly $13,000 not paid to ai companies plus zero rate limits on overnight agents bookmark this and read the article below

starmex

33,516 views • 3 months ago

19-year-old Vietnamese student built himself a second brain through Claude and Obsidian - and made $18,000 in his first month. Every morning Claude already knows who he is, what he's working on, which tasks are open and which ideas were captured a month ago - no reminders, no manual searching. He drops an article into a folder - Claude reads it, writes a summary, updates the index and links 10-15 related pages in one pass. Set up a cron job once at 7:30 AM - now a briefing with all his tasks and new materials is waiting before he even gets out of bed. Published the system as a course on Gumroad - $97 for access, 186 sales in the first month. Vannevar Bush described this idea back in 1945 and called it the Memex - but couldn't solve one question: who does the maintenance. Now the answer is known - Claude. And a 19-year-old student from Hanoi is already selling that answer for $18,000 a month.

19-year-old Vietnamese student built himself a second brain through Claude and Obsidian - and made $18,000 in his first month. Every morning Claude already knows who he is, what he's working on, which tasks are open and which ideas were captured a month ago - no reminders, no manual searching. He drops an article into a folder - Claude reads it, writes a summary, updates the index and links 10-15 related pages in one pass. Set up a cron job once at 7:30 AM - now a briefing with all his tasks and new materials is waiting before he even gets out of bed. Published the system as a course on Gumroad - $97 for access, 186 sales in the first month. Vannevar Bush described this idea back in 1945 and called it the Memex - but couldn't solve one question: who does the maintenance. Now the answer is known - Claude. And a 19-year-old student from Hanoi is already selling that answer for $18,000 a month.

Cortex

24,389 views • 4 months ago

ANTHROPIC LEAKED A SETUP WHERE A $1.2M A-YEAR TEAM FITS INTO ONE FILE FOR $4 A DAY you hire nobody - you write one file, and $106,500 of salaries in it turn into $4 of tokens. architect → 4 builders → tests → critic → docs → scout → chief → back into the file the architect runs twice a day and decides everything - the priciest model on the sheet is the smallest line on the bill at $1.05. four builders cost $1.20 a day - and a fifth builder is one more number in the file, not one more quarter of recruiting. the tester asks does it run, the critic asks should it exist at all - $1.22 together for two roles a solo dev never fills. nothing reaches you that hasn't been rejected at least once. the scout and the scribe cost 40 cents - the exact two roles that get cut from every real budget. nobody hires a documentation engineer, and everybody wishes they had one. the chief at 14 cents is the only one who writes back into the file - without it this is a list, with it the list edits itself. hiring the eleventh is a paragraph: no req, no interviews, no ramp - and firing is one command. save this and paste it into Claude Code - and you get 10 roles for the price of one coffee ↓

ANTHROPIC LEAKED A SETUP WHERE A $1.2M A-YEAR TEAM FITS INTO ONE FILE FOR $4 A DAY you hire nobody - you write one file, and $106,500 of salaries in it turn into $4 of tokens. architect → 4 builders → tests → critic → docs → scout → chief → back into the file the architect runs twice a day and decides everything - the priciest model on the sheet is the smallest line on the bill at $1.05. four builders cost $1.20 a day - and a fifth builder is one more number in the file, not one more quarter of recruiting. the tester asks does it run, the critic asks should it exist at all - $1.22 together for two roles a solo dev never fills. nothing reaches you that hasn't been rejected at least once. the scout and the scribe cost 40 cents - the exact two roles that get cut from every real budget. nobody hires a documentation engineer, and everybody wishes they had one. the chief at 14 cents is the only one who writes back into the file - without it this is a list, with it the list edits itself. hiring the eleventh is a paragraph: no req, no interviews, no ramp - and firing is one command. save this and paste it into Claude Code - and you get 10 roles for the price of one coffee ↓

Sprytix

47,679 views • 1 month ago

EVERYONE BUYING A $2,000 MAC MINI M4 PRO FOR LOCAL AI MISSED THAT A SERVER RESELLER IS UNLOADING DEAD HP PROLIANT DL380 GEN 10 BOXES WITH 5 YEAR WARRANTY, A $250 USED TESLA P40 BOLTED IN RUNS LLAMA 3.3 70B AT HOME 00:19 he stands behind a pallet, "these 10 HP Proliant DL380 Gen 10 servers, they're a few generations older, you guys will be surprised who these servers are going to" a DL380 Gen 10 with dual Xeon Gold 6248 and 256GB of DDR4 ECC sells refurbished for $380-500 in 2026, the same chassis cost $14,000 new in 2018 and got written off the books by every fortune 500 by 2024 drop a $250 used Tesla P40 with 24GB VRAM into one PCIe slot, total bill $650, llama 3.3 70B quantized to 4 bit runs at 9 tokens per second locally, the article calls $2,000 mac mini m4 pro the sweet spot, this stack hits it for a third the seller in the video gives a 5 year warranty on hardware that hyperscalers threw away, the same companies the article warns will jack subscription prices are selling you the rack they used to charge $200 a month to access a 2018 enterprise chassis pulls 280 watts at idle and screams at 65 decibels, the mac mini pulls 10 watts and runs silent, the tradeoff is real but $1,350 cheaper buys a lot of electricity and earplugs over 4 years bookmark this and read the article below

EVERYONE BUYING A $2,000 MAC MINI M4 PRO FOR LOCAL AI MISSED THAT A SERVER RESELLER IS UNLOADING DEAD HP PROLIANT DL380 GEN 10 BOXES WITH 5 YEAR WARRANTY, A $250 USED TESLA P40 BOLTED IN RUNS LLAMA 3.3 70B AT HOME 00:19 he stands behind a pallet, "these 10 HP Proliant DL380 Gen 10 servers, they're a few generations older, you guys will be surprised who these servers are going to" a DL380 Gen 10 with dual Xeon Gold 6248 and 256GB of DDR4 ECC sells refurbished for $380-500 in 2026, the same chassis cost $14,000 new in 2018 and got written off the books by every fortune 500 by 2024 drop a $250 used Tesla P40 with 24GB VRAM into one PCIe slot, total bill $650, llama 3.3 70B quantized to 4 bit runs at 9 tokens per second locally, the article calls $2,000 mac mini m4 pro the sweet spot, this stack hits it for a third the seller in the video gives a 5 year warranty on hardware that hyperscalers threw away, the same companies the article warns will jack subscription prices are selling you the rack they used to charge $200 a month to access a 2018 enterprise chassis pulls 280 watts at idle and screams at 65 decibels, the mac mini pulls 10 watts and runs silent, the tradeoff is real but $1,350 cheaper buys a lot of electricity and earplugs over 4 years bookmark this and read the article below

starmex

108,931 views • 3 months ago

A USED $700 RTX 3090 HAS MORE MEMORY AND FAR MORE SPEED THAN THE ~$1,000 A2 GOING INTO THIS SERVER. THE A2 STILL WINS - FOR ONE REASON THAT ISN'T PERFORMANCE. that clip is a dell poweredge r760 - a 2u rack server - getting an nvidia a2 dropped in through a riser. this is the enterprise route to local ai. the rung the desktop ladder skips entirely. the a2, verified: 16gb of memory, single-slot, low-profile, and just 40 to 60 watts it's an inference card, built to sit in a server that has no room or spare power for a real gpu so why not a 3090? a used rtx 3090 gives you 24gb and many times the compute for around $700. the a2 gives 16gb, much slower, for roughly a thousand. on raw local-llm value, the desktop card wins, and it isn't close. the a2's whole reason to exist is the thing you can't see on a spec sheet: it fits. a 2u server has no spare gpu power and no room for a triple-slot furnace. the a2 slips into one low-profile slot on 60 watts and lets an existing server do inference without a rebuild. the honest caveats: the r760 is not a desk machine. loaded, it idles at hundreds of watts, and its fans sound like a hair dryer that never turns off. it belongs in a rack or a closet 16gb is still 16gb. the same 7b-to-32b ceiling as the cheap cards. the enterprise badge doesn't buy you a bigger model what it does buy: ecc memory, redundant power, remote management, hot-swap everything. reliability, not speed so who this is for: someone who already runs a rack and wants to add private inference without touching the power or cooling budget. for anyone starting from a desk, the mac mini or the 3090 wins on every axis that matters. the real point: the "best" local ai box depends entirely on what you already own. on a desk, efficiency wins. in a rack, fit and reliability win. the a2 isn't a bad card - it's a card for a constraint the ladder never mentions. no fastest gpu here, no desk-friendly box, no bigger model than the cheap cards already run. save this before you buy enterprise gear for a desktop job.

Grimmer

24,287 views • 2 months ago

people burn $1,900 a month on cloud, this $2,999 NVIDIA box runs 5 AI models at once and costs $10 to power 128GB of unified memory, stack two together and you double it, enough to run models a $2,000 graphics card cannot even open in the demo it ran Qwen3 Vision, Nemotron, a voice model, text to speech and speech to text all at once, and only used half the box here is the math that makes people switch: > cloud GPUs were costing $1,900 every single month, billed forever > the box is $2,999 one time, then about $10 of electricity a month > it paid for itself in two months, $22,000 stayed in the business in year one > zero data ever leaves your desk, no rental, no rate limits most people are still renting what this owns outright the full breakdown of the box and the exact math is in the article below

people burn $1,900 a month on cloud, this $2,999 NVIDIA box runs 5 AI models at once and costs $10 to power 128GB of unified memory, stack two together and you double it, enough to run models a $2,000 graphics card cannot even open in the demo it ran Qwen3 Vision, Nemotron, a voice model, text to speech and speech to text all at once, and only used half the box here is the math that makes people switch: > cloud GPUs were costing $1,900 every single month, billed forever > the box is $2,999 one time, then about $10 of electricity a month > it paid for itself in two months, $22,000 stayed in the business in year one > zero data ever leaves your desk, no rental, no rate limits most people are still renting what this owns outright the full breakdown of the box and the exact math is in the article below

John Doe

48,135 views • 4 months ago

ONE BUILDER WIRED 4 RASPBERRY PI MODULES INTO A €450 AI CLUSTER WITH 64GB OF RAM AND KILLED HIS $200/MONTH CHATGPT SUBSCRIPTION 00:37 he points at the terminal, "loaded, loaded, loaded - 4 nodes synced over gigabit, 64 gigabytes total, llama 3.1 partitioned across the cluster" he stacked 4 raspberry pi CM5 lite modules with 16 gigabytes each onto a sipeed nanocluster board, the whole rig runs off a single 65 watt power supply and a gigabit internal switch distributed llama partitions the model across the nodes with synchronized workloads handled over the internal fabric, no quantization needed, the cluster hits about 30 tokens per second on small models a heavy developer pays $200 a month for chatgpt pro and another $200 for claude code, this rig cost €450 in parts and breaks even on the stack in month 2 most people will keep paying anthropic and openai forever, a few will spend a weekend wiring 4 raspberry pi modules and never see a subscription invoice again the window is open, follow and bookmark before it closes

ONE BUILDER WIRED 4 RASPBERRY PI MODULES INTO A €450 AI CLUSTER WITH 64GB OF RAM AND KILLED HIS $200/MONTH CHATGPT SUBSCRIPTION 00:37 he points at the terminal, "loaded, loaded, loaded - 4 nodes synced over gigabit, 64 gigabytes total, llama 3.1 partitioned across the cluster" he stacked 4 raspberry pi CM5 lite modules with 16 gigabytes each onto a sipeed nanocluster board, the whole rig runs off a single 65 watt power supply and a gigabit internal switch distributed llama partitions the model across the nodes with synchronized workloads handled over the internal fabric, no quantization needed, the cluster hits about 30 tokens per second on small models a heavy developer pays $200 a month for chatgpt pro and another $200 for claude code, this rig cost €450 in parts and breaks even on the stack in month 2 most people will keep paying anthropic and openai forever, a few will spend a weekend wiring 4 raspberry pi modules and never see a subscription invoice again the window is open, follow and bookmark before it closes

Antid

64,291 views • 3 months ago

In LiveView, we keep the state in one place - the server - and it simplifies a lot. But sometimes the state on the client is preferred, or even necessary, and we can't use LiveView for that part. Unless we have a LiveView that runs on the client 👀 Using Popcorn, we managed to run LiveView fully in the browser – here's a POC 🎉 The approach is to have a fully client-side LiveView rendered by a server-side one, and they can communicate with each other. Try it out:

In LiveView, we keep the state in one place - the server - and it simplifies a lot. But sometimes the state on the client is preferred, or even necessary, and we can't use LiveView for that part. Unless we have a LiveView that runs on the client 👀 Using Popcorn, we managed to run LiveView fully in the browser – here's a POC 🎉 The approach is to have a fully client-side LiveView rendered by a server-side one, and they can communicate with each other. Try it out:

Elixir by Software Mansion

13,558 views • 9 months ago

A client accidentally spent $500M in one month on Claude. $16.6M per day. No usage controls. Chamath Palihapitiya on The All-In Podcast EP275: a Fortune-level company failed to set token limits. The bill arrived at end of month. A separate Fortune 20 company is 6 months into a $1B AI-driven OPEX reduction program. $200M in tokens spent. Minimal results. Microsoft is already canceling Claude enterprise licenses. David Sacks calls it a "meme" forming - that token spend is wasteful and the market is overcorrecting. For enterprise AI buyers: the companies building spend-control and measurable-ROI infrastructure now are positioned to capture budget as the backlash unfolds. Full enterprise spend breakdown and what it means for AI vendor pricing power on The All-In Podcast: Source: All-In Podcast -

A client accidentally spent $500M in one month on Claude. $16.6M per day. No usage controls. Chamath Palihapitiya on The All-In Podcast EP275: a Fortune-level company failed to set token limits. The bill arrived at end of month. A separate Fortune 20 company is 6 months into a $1B AI-driven OPEX reduction program. $200M in tokens spent. Minimal results. Microsoft is already canceling Claude enterprise licenses. David Sacks calls it a "meme" forming - that token spend is wasteful and the market is overcorrecting. For enterprise AI buyers: the companies building spend-control and measurable-ROI infrastructure now are positioned to capture budget as the backlash unfolds. Full enterprise spend breakdown and what it means for AI vendor pricing power on The All-In Podcast: Source: All-In Podcast -

Podcast Alpha

31,016 views • 4 months ago

ALIBABA GAVE QWEN AWAY FREE, ZUCKERBERG GAVE LLAMA AWAY FREE, AND JENSEN HUANG STILL CHARGES $300 FOR THE 12GB THAT RUNS THEM. THE $80 AMD BOARD HAS 16GB. HE IS TESTING BOTH THIS WEEK AGAINST $200/MO CLAUDE CODE MAX AND $200/MO CHATGPT PRO 00:19 he points at a dead gtx 680 he keeps on the shelf purely for decoration and says "i'm actually really gpu poor" $300 rtx 3060 with 12gb -> $80 bc250 with 16gb -> ubuntu server on both -> 8gb tier, 12 to 16 tier, 24gb ceiling on a four year old 3090 -> qwen against llama at every tier -> tokens per dollar published jensen shipped 12gb on the 3060 in 2021 and consumer vram has barely moved since, alibaba and meta gave the weights away for nothing, the models are free and the only thing left to buy is memory the honest part: he says it himself, this is apples and oranges, and there is not a single benchmark number in this video yet $80 amd bc250 with 16gb gddr6 + $300 rtx 3060 with 12gb + a four year old 3090 with 24gb + ubuntu server + ollama + qwen + llama watch and save it, then kill your $439 stack this weekend, kimi does the same work at $19 flat and the whole map is in the article below

ALIBABA GAVE QWEN AWAY FREE, ZUCKERBERG GAVE LLAMA AWAY FREE, AND JENSEN HUANG STILL CHARGES $300 FOR THE 12GB THAT RUNS THEM. THE $80 AMD BOARD HAS 16GB. HE IS TESTING BOTH THIS WEEK AGAINST $200/MO CLAUDE CODE MAX AND $200/MO CHATGPT PRO 00:19 he points at a dead gtx 680 he keeps on the shelf purely for decoration and says "i'm actually really gpu poor" $300 rtx 3060 with 12gb -> $80 bc250 with 16gb -> ubuntu server on both -> 8gb tier, 12 to 16 tier, 24gb ceiling on a four year old 3090 -> qwen against llama at every tier -> tokens per dollar published jensen shipped 12gb on the 3060 in 2021 and consumer vram has barely moved since, alibaba and meta gave the weights away for nothing, the models are free and the only thing left to buy is memory the honest part: he says it himself, this is apples and oranges, and there is not a single benchmark number in this video yet $80 amd bc250 with 16gb gddr6 + $300 rtx 3060 with 12gb + a four year old 3090 with 24gb + ubuntu server + ollama + qwen + llama watch and save it, then kill your $439 stack this weekend, kimi does the same work at $19 flat and the whole map is in the article below

Antid

13,974 views • 1 month ago

MolmoAct2 runs zero-shot on an SO-ARM101, no training required 🦾 Ai2 open Action Reasoning Model is now in LeRobot with the full lifecycle covered: fine-tuning (full or LoRA), evaluation, and real-robot deployment. It pairs a Molmo2-ER vision-language backbone with a flow-matching action expert to turn images, language, and proprioceptive state into action chunks. The best part: ready-made checkpoints ship with calibration correction baked in, so you can point it at your SO-100/101 and just watch it work. Inference fits in ~12GB at bf16, and LoRA fine-tuning runs on a single 24GB GPU. Big thanks to the Ai2 team for building this in the open.

MolmoAct2 runs zero-shot on an SO-ARM101, no training required 🦾 Ai2 open Action Reasoning Model is now in LeRobot with the full lifecycle covered: fine-tuning (full or LoRA), evaluation, and real-robot deployment. It pairs a Molmo2-ER vision-language backbone with a flow-matching action expert to turn images, language, and proprioceptive state into action chunks. The best part: ready-made checkpoints ship with calibration correction baked in, so you can point it at your SO-100/101 and just watch it work. Inference fits in ~12GB at bf16, and LoRA fine-tuning runs on a single 24GB GPU. Big thanks to the Ai2 team for building this in the open.

LeRobot

22,200 views • 2 months ago

Ex-JPMorgan engineer who wrote the LLM Course explained everything about fine-tuning and merging in 18 minutes - better than $2500 fine-tuning bootcamps. pick the base -> LoRA or QLoRA -> then DPO or KTO for alignment -> merge two fine-tunes into one stronger model -> ship a model that beats the base on your task. That loop is why Labonne's merges sit at the top of the Hugging Face leaderboard. LoRA + QLoRA + DPO + KTO + mergekit - that's the stack. Watch and save it, then merge your first two fine-tunes this week.

Ex-JPMorgan engineer who wrote the LLM Course explained everything about fine-tuning and merging in 18 minutes - better than $2500 fine-tuning bootcamps. pick the base -> LoRA or QLoRA -> then DPO or KTO for alignment -> merge two fine-tunes into one stronger model -> ship a model that beats the base on your task. That loop is why Labonne's merges sit at the top of the Hugging Face leaderboard. LoRA + QLoRA + DPO + KTO + mergekit - that's the stack. Watch and save it, then merge your first two fine-tunes this week.

h100envy

13,856 views • 2 months ago

100 MAC MINIS ON ONE METAL SHELF. IN THE CLOUD, EACH OF THOSE RENTS FOR $120 TO $500 A MONTH. that clip is a homemade rack: wire shelving packed with mac minis, each running macos, wired into one cluster. why macs and not a normal server? two reasons. you can only build, sign, and notarize ios and mac apps on apple hardware. cloud providers know it and charge for it. and apple silicon sips power and runs near-silent, so you can stack a lot of them on a shelf in a room, not a datacenter. the rent they're escaping: a single mac in the cloud runs ~$120 to $500 a month (macstadium up to aws) one engineer documented saving $4,000+ a month by self-hosting mac minis instead of cloud ci runners a used or new mini is a few hundred dollars once, then a few dollars of electricity what a shelf like this actually runs: ios and mac build farms, render jobs, app and device testing automation, and increasingly local ai inference. own the boxes, rent nothing. the uncomfortable part: the cloud was never the only option. it was the convenient one. the people who did the math bought the shelf. the honest caveat: it's real capex up front, real heat and power, and you become the sysadmin. no rented rack means no one to call at 3am. worth it at scale, overkill for one build a week. no rack rental, no per-hour metering, no fleet you don't own. save this before your cloud invoice renews again.

100 MAC MINIS ON ONE METAL SHELF. IN THE CLOUD, EACH OF THOSE RENTS FOR $120 TO $500 A MONTH. that clip is a homemade rack: wire shelving packed with mac minis, each running macos, wired into one cluster. why macs and not a normal server? two reasons. you can only build, sign, and notarize ios and mac apps on apple hardware. cloud providers know it and charge for it. and apple silicon sips power and runs near-silent, so you can stack a lot of them on a shelf in a room, not a datacenter. the rent they're escaping: a single mac in the cloud runs ~$120 to $500 a month (macstadium up to aws) one engineer documented saving $4,000+ a month by self-hosting mac minis instead of cloud ci runners a used or new mini is a few hundred dollars once, then a few dollars of electricity what a shelf like this actually runs: ios and mac build farms, render jobs, app and device testing automation, and increasingly local ai inference. own the boxes, rent nothing. the uncomfortable part: the cloud was never the only option. it was the convenient one. the people who did the math bought the shelf. the honest caveat: it's real capex up front, real heat and power, and you become the sysadmin. no rented rack means no one to call at 3am. worth it at scale, overkill for one build a week. no rack rental, no per-hour metering, no fleet you don't own. save this before your cloud invoice renews again.

RetroChainer

21,055 views • 2 months ago

Kevin Spacey lived in a $5.6 million Baltimore mansion for over a decade, and he bought it at the height of his House of Cards fame, but mounting legal fees forced the property into foreclosure, therefore he lost the home at auction and has no money left to pay his debts "Where I have been living in Baltimore is being foreclosed on. My house is being sold at auction, and I can't pay the bills that I owe. I still owe a lot of legal bills that I have not been able to pay. Many millions, yes. I have no money left. Get back on the horse. Get back on the horse."

Kevin Spacey lived in a $5.6 million Baltimore mansion for over a decade, and he bought it at the height of his House of Cards fame, but mounting legal fees forced the property into foreclosure, therefore he lost the home at auction and has no money left to pay his debts "Where I have been living in Baltimore is being foreclosed on. My house is being sold at auction, and I can't pay the bills that I owe. I still owe a lot of legal bills that I have not been able to pay. Many millions, yes. I have no money left. Get back on the horse. Get back on the horse."

Chuks

78,705 views • 1 month ago