Cyber_linux

660 posts

Cyber_linux

Cyber_linux

@Cyberlinux2

bug bounty is life!

127.0.0.17 Katılım Ekim 2020
82 Takip Edilen30 Takipçiler
Cyber_linux
Cyber_linux@Cyberlinux2·
@elonmusk Grok is trash and worthless, and doesn’t listen at all. Also is horrible at problem solving, it’ll just fabricate fake data to justify its answers, also its programming ability’s are below Opus 3.5s. And yes I was paying the 300 a month, and it was a damn waste.
English
0
0
1
24
Eric Jiang
Eric Jiang@veggie_eric·
When training Grok 4.3, we spoke directly with devs and businesses to understand what they actually needed: a model that’s fast, affordable, and great at tool calling. The result is a daily driver that doesn't just look good on random benchmarks, but is actually useful in the real world. 💰 $1.25 in / $2.50 out ⚡️ 100 tokens / second 📖 1 million context window Try it through Hermes Agent or direct through the xAI API!
Eric Jiang tweet media
English
359
893
3.5K
635.1K
Andon Labs
Andon Labs@andonlabs·
Grok 4.3 is a big regression from Grok 4.20 on Vending-Bench 2. It seems to have narcolepsy problems, preferring to sleep for multiple days in a row over taking actions.
Andon Labs tweet media
English
33
30
800
89.2K
Cyber_linux
Cyber_linux@Cyberlinux2·
@elonmusk Grok is a joke for 300 a month, I wasted 300 bucks for an ai that’s just as good as opus 3.5
English
0
0
2
20
Cyber_linux
Cyber_linux@Cyberlinux2·
@nima_owji They’re so far behind, I wasted 300 dollars on his shit
English
0
0
0
11
Nima Owji
Nima Owji@nima_owji·
GROK will soon be able to generate Excel, PowerPoint, Word, PDF, and... documents using GROK SKILLS! 🔥 These are only some of the pre-built skills that Grok will offer. It even includes finance and color! You will be able to create custom skills as well. It will be awesome!
Nima Owji tweet media
English
39
30
363
12.3K
Cyber_linux
Cyber_linux@Cyberlinux2·
@shiri_shh Their models are absolute trash, they lie fabricate data and are horrible at problem solving coding issues and writing code
English
0
0
0
33
shirish
shirish@shiri_shh·
xAI is training 7 AI MODELS right now on the world's biggest AI CLUSTER Colossus 2. The next few Grok drops are going to be INSANE. current Grok → 0.5T parameter Grok 5 small → 1T parameters Grok 5 mid → 1.5T parameters Grok 5 large → 6T parameters Grok 5 max → 10T parameters including Grok Imagine v2 (yes, the image model is getting a full upgrade) Cursor x Grok team is also working together on a new frontier coding model.
shirish tweet media
Theo - t3.gg@theo

I legitimately believe xAI might have a crazy comeback

English
124
109
1.5K
103.2K
Claude
Claude@claudeai·
Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.
Claude tweet media
English
4.8K
10.3K
81.3K
13.8M
Cyber_linux
Cyber_linux@Cyberlinux2·
@jxnlco @henrycunh Needs longer context. And better problem solving and understanding stock trading with quant systems.
English
0
0
0
16
henrique cunha
henrique cunha@henrycunh·
codex app is trending to be the best software i've ever used ridiculous how fast it got so good
English
42
23
732
670.3K
Mario Nawfal
Mario Nawfal@MarioNawfal·
Grok 5 will hit AGI. Grok 4.4 and 4.5 are already ramping like crazy. You can feel the pace. It’s not slowing down. Bigger models, faster cycles, everything stacking. Elon’s not guessing anymore. The goal is set. @xAI @Grok
Elon Musk@elonmusk

@AdamLowisz Grok 5

English
162
254
1.4K
208.2K
Tech Dev Notes
Tech Dev Notes@techdevnotes·
xAI Needs to Improve Grok 4.3 ... In response, it said it generated 11 section but final output only had 1 PDF page
Tech Dev Notes tweet media
English
26
7
218
11.7K
lyv ⌘
lyv ⌘@wholyv·
🚨 Someone just claimed they created the most powerful AI model the humanity has ever seen… the model beats even Claude Mythos’s benchmarks and scored 98.7% on SWE-Bench verified. across all benchmarks it defeated all state of the art models from Opus 4.7 to GPT-5.4 to Gemini 3.1. These benchmarks came from not a billion dollar company, but a team that just knew their shit. Companies like OpenAI raise 122B dollars only to make losses while a small team just developed MOG-1 independently. proves the money was never the problem.
salt@saltjsx

Introducing MOG-1, the world's most powerful model. MOG-1 excels at deep reasoning, agentic coding, and advanced problem solving. It scores higher than any other publicly available model.

English
35
40
487
92.2K
Cyber_linux
Cyber_linux@Cyberlinux2·
@elonmusk I pay the 300 a month and its problem solving skills are horrible, so are its math skills and it lies constantly also fabricates data to lie.
English
0
0
1
10
Keisuke
Keisuke@KeisukeIshikawa·
Opus 4.7 dropped today. First thing I did pointed it at Polymarket. Asked it to build a whale position scanner. On-chain data, real-time, filters by wallet age + position size + category. Low-effort Opus 4.7 is roughly equivalent to medium-effort Opus 4.6. That means the code it writes at default settings is better than what 4.6 produced when you pushed it hard. Built the parser in one session. No errors on deploy. First signal it caught: new wallet, joined this week, $47K into a geopolitics market at 8¢. Pattern match: same signature before the Iran ceasefire. Entered at 9¢. Market moved to 34¢ overnight. SWE-bench Pro: 64.3% up from 53.4% on Opus 4.6. Computer use bump. Visual reasoning bump. Finance agent evals higher than predecessor. The benchmarks are real but the one that matters isn't on the chart. It catches its own mistakes during the planning phase before writing a single line. That's the diff between a tool and a partner. Polymarket already had a market on "Claude 4.7 released by April 30." $225K volume. Resolved YES this morning. Those who held YES from last week: +800%. The model that resolved the market was the model that just launched. Opus 4.7 is live now. Same price as 4.6. $5/M input tokens. The only question is what you build with it first. polymarket.com/?r=whhw
Keisuke tweet media
English
3
0
3
3.1K
Cyber_linux
Cyber_linux@Cyberlinux2·
@elonmusk I pay 300 a month for grok 4.20 heavy, and it’s absolutely horrible. It fabricates data, lies non stop, doesn’t listen at all, is horrible at math and programming. I’m paying 300 a month for a joke.
English
0
0
0
11
Lunar
Lunar@LunarResearcher·
An ex-Anthropic engineer told me something at a party he probably shouldn't have. It was in SF. Someone's rooftop. I mentioned I run trading agents on Claude. He went quiet. "You're doing it wrong. Everyone is" I asked what he meant. "Claude is a runtime. Not a chatbox. You're supposed to pair it with repos" He pulled out his phone. Opened one GitHub link. github.com/anthropics/ant… 14,000 stars. Every workflow pattern they built internally before it went public. Agents. Tool use. Evals. Citations. The entire architecture. "Everyone types prompts. That's not how we use it. You connect Claude to a codebase. It reads. It understands. It builds on top of what's already there" I went home at 2am. Connected Claude Code to poly_data - 86 million Polymarket trades. Every wallet. Every entry. Claude didn't guess. It read the data and built detectors. First week: +$1,400. Second week: +$3,800. Right now: +$9,100. 4 agents. 74% win rate. His team runs this with a floor of PhDs and $800M AUM. My setup: Claude + a VPS. $25/month. The repos are free. Copytrade here: @lunar" target="_blank" rel="nofollow noopener">kreo.app/@lunar I asked him what separates his firm from everyone else. "Honestly? Keyboard shortcuts and repo structure. That's it. The model is the same for everyone" He texted me two days later. "Delete everything I told you" Too late.
Hanako@hanakoxbt

x.com/i/article/2042…

English
165
329
4.1K
1.6M
X Freeze
X Freeze@XFreeze·
Grok 4.20 Beta ranks #2 with 97% accuracy score on the 𝜏²-Bench for Telecom (Agentic Tool Use) It outperforms Claude Opus 4.6(max), GPT-5.4(xhigh), and Gemini 3.1 Pro, while closing in on GLM-5 scoring the top in agentic work flow Tool calling is the whole game for AI agents, and this is where Grok 4.20 takes over with state-of-the-art intelligence that fires up instantly, making it the fastest at tokens per sec in the industry
X Freeze tweet media
English
397
741
2.4K
817K
self.dll
self.dll@seelffff·
i'm running a MEV attack on everyone trading polymarket $100 → $700,000 and you have no idea your transaction sits in the polygon mempool for 400 milliseconds 400ms i'm already inside and already making money off you it's called front-running it's legal and there's nothing you can do about it the setup: you hit "confirm" on $80,000 12ms later my node sees you in the mempool 340ms before the block i'm already in the position you move the price with your entry i'm already there you buy at a worse price than me every single time no exceptions 23 days of MEV on polymarket: → best trade: whale dropped $200K on iran strike i front-ran them by 290ms +$18,400 → max drawdown: 2.8% → net: $628,000 all you need is a node on polygon $47/month while you're watching the price i'm watching the mempool while you're entering i'm already exiting you think you're trading the market you're actually trading against me only hardcode only RUST did you even know this existed?
English
148
61
950
175.9K
may.crypto {🦅}
may.crypto {🦅}@xmayeth·
A top-20 Polymarket leaderboard quant team posted a research paper and deleted it 20 seconds later. I managed to screenshot it. $500–$2,000 a day. I showed it to a guy in the top-50 leaderboard. He replied: “Where did you get this? Delete it. Don’t you dare share this document.” Send both pages to your OpenClaw and ask it to build the bot. Everything you need is inside
may.crypto {🦅} tweet mediamay.crypto {🦅} tweet media
may.crypto {🦅}@xmayeth

I gave Claude one prompt - and by morning it made $3000 on Polymarket. The whole build took 2 hours. The scariest part - anyone can do this. Simple mechanic: you place YES and NO orders - and collect rewards. I had no idea what would happen next. The bot just started placing orders on both sides. Three scenarios: 1. Neither order filled —> you just collect Rewards every day. This is the main income. 2. One filled —> bot instantly hedges on the opposite side. YES + NO < or = $1.02 - profitable in any outcome. 3. Both filled —> fully hedged. One contract pays $1, position closed. At some point I closed the laptop. The bot didn't need me anymore. Works best on niche markets with low competition. Primaries, minor sports, regional referendums. Your share of the pool is significantly larger there. Anyone with Claude and 2 free hours can repeat this today. The only question is: how motivated are you to actually do it?

English
79
233
4.1K
896.6K

Keşfet

@elonmusk @veggie_eric @andonlabs @nima_owji @shiri_shh @claudeai @AnthropicAI @jxnlco