Qwinah

348 posts

Qwinah banner
Qwinah

Qwinah

@MaaSonder

Sharing what really happened with AI. Save Claude Tokens : https://t.co/MBdCH7yJbC

เข้าร่วม Haziran 2025
46 กำลังติดตาม37 ผู้ติดตาม
0xMarioNawfal
0xMarioNawfal@RoundtableSpace·
THE US GOVERNMENT SHUT DOWN AN AI MODEL OVERNIGHT AND MILLIONS LOST ACCESS INSTANTLY Claude Code runs from your terminal, plugs into VS Code, JetBrains and Slack, and Agent OS lets you switch models in one workflow. One gets pulled, you flip to another and keep building.
English
10
6
81
40.7K
Qwinah
Qwinah@MaaSonder·
@SBlockspy For sure looks to be one to watch
English
0
0
1
17
SBlockSpy
SBlockSpy@SBlockspy·
Tron Inc. now holds over 700.4M TRX after adding another 157,392 TRX to its treasury...thats like someone already owning a mountain of pizza and still grabbing 1 more slice 😅 Small buy, big signal: they're still accumulating, not slowing down...long-term convction always gets my attention🚀 #TRX
SBlockSpy tweet media
English
1
0
3
61
Prasenjit
Prasenjit@Star_Knight12·
asked fable 5 one last question before it got taken away
Prasenjit tweet media
English
10
2
68
3.7K
Qwinah
Qwinah@MaaSonder·
Been using the @garrytan “Boil The Ocean” prompt and it’s something else. The premise of the prompt is that the difference of AI completing half the job, versus the entire thing is like 3 droplets of water. So Gary’s point was to do the whole thing, never talk back to me and make no mistakes “Boil The Ocean”. Let me know how it goes. Prompt⤵️ Remember when implementing: The marginal cost of completeness is near zero with AI. Do the whole thing. Do it right. Do it with tests. Do it with documentation. Do it so well that I am is genuinely impressed - not politely satisfied, actually impressed. Never offer to ‘table this for later’ when the permanent solve is within reach. Never leave a dangling thread when tying it off takes five more minutes. Never present a workaround when the real fix exists. The standard isn’t ‘good enough’ - it’s ‘holy shit, that’s done.’ Search before building. Test before shipping. Ship the complete thing. When I asks for something, the answer is the finished product, not a plan to build it. Time is not an excuse. Fatigue is not an excuse. Complexity is not an excuse. Boil the ocean.
Qwinah tweet media
English
1
0
3
50
Qwinah
Qwinah@MaaSonder·
@RoundtableSpace Good morning mate, another day to reminisce on Fable 5 games 🥲
English
0
0
0
58
0xMarioNawfal
0xMarioNawfal@RoundtableSpace·
GM CT☀️ ENJOY YOUR DAY☕
0xMarioNawfal tweet media
English
121
2
164
44.4K
Qwinah
Qwinah@MaaSonder·
@Adidotdev Yea been seeing a lot about GLM5.2 gonna give it a go
English
0
0
1
149
Adit_Yah ☄️
Adit_Yah ☄️@Adidotdev·
Codex 10x limits gone Opus 4.8 self fixes bugs Fable 5 banned by the government GPT-5.6 still "coming soon" this week broke me.
English
16
0
121
4.5K
Qwinah
Qwinah@MaaSonder·
@Adidotdev Yea bro, has me just sitting here now not knowing what to do without Fable 🥲
English
1
0
0
26
Qwinah
Qwinah@MaaSonder·
@VadimStrizheus Your telling me all those knock off games wasn’t AGI ?!
English
0
0
0
115
Vadim
Vadim@VadimStrizheus·
Why is everybody acting that Claude Fable 5 was AGI??
English
13
0
41
2.6K
Qwinah
Qwinah@MaaSonder·
@Kierbnak @RoundtableSpace I mean it can basically do anything you put in opus which will then orchestrate it across all the other models
English
0
0
1
17
0xMarioNawfal
0xMarioNawfal@RoundtableSpace·
CORPORATIONS ARE PAYING ONE PERSON $15,000 A PROJECT AND HE DELIVERS IN 72 HOURS He routes tasks across multiple AI models at once. Opus 4.8 orchestrates, GPT-5.5 handles copy, Gemini processes multimodal, and parallel subagents make sure nothing slows down. Studios are still hiring coordinators to manage the pipeline. He automated it and became the studio. Over $40,000 a month from three clients and one laptop.
English
29
18
198
48.6K
Gera92
Gera92@SolG_420·
@RoundtableSpace three clients and a laptop. i cant even get one client to respond to my emails.
English
1
0
1
124
Qwinah
Qwinah@MaaSonder·
Fable 5 shut down complete timeline‼️ Anthropic released Fable 5 June 9th. Soon after, US authorities raised national security concerns around potential jailbreak misuse and cyber capability risks. A US export control directive then required Anthropic to restrict access for foreign nationals. The problem being they couldn’t cleanly separate users by region or nationality without breaking the system. So Anthropic did the only workable move. They shut down Fable 5. There’s still no public proof of a major new breakthrough exploit. The risk framing is based on limited disclosed security concerns, not a clearly demonstrated catastrophic failure. This just tells us that frontier AI is starting to be treated like export-controlled infrastructure (similar to chips or encryption). And when that happens, product access becomes policy driven, not company driven. I want to clarify that this wasn’t a model being “too dangerous.” It was a policy + technical enforcement clash that forced a full shutdown.
Qwinah tweet media
English
1
0
1
89
Qwinah
Qwinah@MaaSonder·
@ns123abc They knew Anthropic couldn’t get it done in 90minutes
English
1
0
0
928
NIK
NIK@ns123abc·
🚨BREAKING: The U.S government gave Anthropic 90 minutes to shut down Fable and Mythos “Amazon AND others” called senior administration officials to warn about models’ capabilities Then: 1:00pm: Government calls. “Take it down.” Cites “national security threat.” No details. Anthropic asks what the threat is so they can fix it. Government said NO. 5:30pm: Commerce letter arrives with export controls. You have 90 minutes…
NIK tweet mediaNIK tweet mediaNIK tweet media
English
274
409
4K
599.8K
Macro Bombastic
Macro Bombastic@MacroBombastic·
@RoundtableSpace Crazy how much time this saves, but you still need a good eye for design to make it actually work
English
1
0
1
57
0xMarioNawfal
0xMarioNawfal@RoundtableSpace·
FIGMA AND CLAUDE CODE JUST KILLED THE ENTIRE LANDING PAGE REBUILD WORKFLOW Someone built a pipeline that captures any competitor's page through Figma's new Chrome extension and hands it to Claude Fable 5. It comes back rebuilt in your brand. Your copy, your fonts, your images, your voice. Start to finish in about 15 minutes and $2 in API costs.
English
16
6
107
49.1K
Qwinah
Qwinah@MaaSonder·
@0xrinx No, not really. I mean it still needs heavy external tools to work well. An incredible leap nonetheless but not AGI
English
0
0
2
51
RinX 🪐
RinX 🪐@0xrinx·
is fable 5 the moment people look back on and say: "that was the beginning of AGI"?
English
14
2
36
1.9K
GREG ISENBERG
GREG ISENBERG@gregisenberg·
Fable is banned. Long live local AI. Full episode breaking down exactly how to get good at local models. the runtime, the hardware, quantization, connecting it to Hermes agent and local AI startup ideas (25 minutes)
GREG ISENBERG@gregisenberg

The takeaway from Fable 5 being BANNED by the government: GET GOOD AT LOCAL MODELS SO YOU HAVE 100% CONTROL. My entire weekend was going to be building my craziest ideas with Fable 5. That's now cancelled. So instead of building with Fable this weekend, I've decided I'll go deep on local models: 1. Start with the runtime. Download Ollama or LM Studio first. This is the thing that actually runs models on your machine. 2. Match the model to your hardware. A model's size is measured in billions of parameters (7B, 32B, 70B). Bigger is smarter but needs more memory. Rule of thumb: a 7B model runs on almost any laptop, a 32B needs a good Mac with 32GB+ RAM, a 70B needs serious hardware like a DGX Spark or a maxed-out Mac Studio. 3. Know which model for which job. Qwen 3 is the best all-around choice for most tasks. DeepSeek for reasoning and coding. Gemma 4 when you need something tiny that runs on a phone. Llama when you want the biggest community and the most fine-tunes. 4. Quantization. You can shrink a model to run on weaker hardware with barely any quality loss. Look for versions labeled Q4 or Q5. This is how a model that "needs" a server runs on your laptop. Learning this one concept changes everything. 5. Connect it to your agent. Point Hermes or your agent stack at a local model. 6. Context window is your real constraint locally. Cloud models give you huge context for free. Local models make you pay for it in memory. A bigger context window eats RAM fast. Keep your sessions tight and your prompts lean or your machine chokes. 7. Learn to give local models tools. A smaller local model with web search, file access, and code execution beats a giant model with none. The capability gap closes fast when you wire up the right tools. The model is the engine but the tools are the wheels. 8. Fine-tuning is more accessible than you think. You don't need this on day one, but know it exists. You can take an open model and train it on your own data so it gets good at your specific domain. I'll probably do a breakdown at some point on this @startupideaspod if people are into it. The lesson from this ban is basically don't build your entire workflow on something that can disappear with a single letter. Own part of your stack. Local models are insurance. It reminds me when people realized they don't own social media accounts. And then you saw people build email lists etc. I remember running a startup and my biggest traffic source was organic FB. All of a sudden, algo changed, and I lost 99% of my traffic. Same sorta moment (but bigger) for AI. This is a wake up call.

English
62
89
691
90.5K