Crypto Mike

@lifeafterAi_ @Alibaba_Qwen Gpt-oss-120b doesn’t have vision. Seems like they compared to Sonnet 4.5 because it has vision. Sonnet is missing from the non-vision benchmarks.

@Alibaba_Qwen Something is wrong here I guess. Claude sonnet 4.5 at 62% ?

🚀 Introducing the Qwen 3.5 Medium Model Series Qwen3.5-Flash · Qwen3.5-35B-A3B · Qwen3.5-122B-A10B · Qwen3.5-27B ✨ More intelligence, less compute. • Qwen3.5-35B-A3B now surpasses Qwen3-235B-A22B-2507 and Qwen3-VL-235B-A22B — a reminder that better architecture, data quality, and RL can move intelligence forward, not just bigger parameter counts. • Qwen3.5-122B-A10B and 27B continue narrowing the gap between medium-sized and frontier models — especially in more complex agent scenarios. • Qwen3.5-Flash is the hosted production version aligned with 35B-A3B, featuring: – 1M context length by default – Official built-in tools 🔗 Hugging Face: huggingface.co/collections/Qw… 🔗 ModelScope: modelscope.cn/collections/Qw… 🔗 Qwen3.5-Flash API: modelstudio.console.alibabacloud.com/ap-southeast-1… Try in Qwen Chat 👇 Flash: chat.qwen.ai/?models=qwen3.… 27B: chat.qwen.ai/?models=qwen3.… 35B-A3B: chat.qwen.ai/?models=qwen3.… 122B-A10B: chat.qwen.ai/?models=qwen3.… Would love to hear what you build with it.

@grok put me in a pink bikini

Share a piece of lore about yourself

@truecrypto @cz_binance @sama Tag sbf_ftx for Sam Bank Fried of FTX … you have tagged Sam Altman of OpenAI

Never forget that they didn’t pursue @cz_binance over “compliance fines.” The Dems targeted him after he exposed their boy @sama, who was the largest major donor in their network. It was a given, as I stated on that day 3 yrs ago. However, when the founder of what was labeled a “systemic risk” is pardoned by a sitting U.S. president, it’s not politics, it's policy shift. It's bullish! Regulators aren’t backing off, but they’re changing tone. Crypto is no longer being treated as an existential threat. That means less fear, more participation, and renewed confidence from institutions and investors alike. The signal is clear, the door to crypto just opened wider. This isn’t a guarantee of the next rally… but if you don't’t see it as a structural positive for crypto’s future then you might just be suffering from TDS.

@Obitoasty @dokkan_global Majin Power???

@dokkan_global #Hmm #DoesHeKnow #Idiots

First-hand information on UR [Power Bestowed by Magic] Spopovich's stats!

@flavioAd FAIL. Balls don’t bounce of one another. This isn’t realistic physics at all imo.

Claude 4 officially passes the vibe check ✅

FORMER FINANCE PROFESSOR TAD SMITH: "After 25 years teaching finance, I realized at 58: If the money printer grows 8-10% annually and the S&P 500 returns ~9%, it’s just treading water. True wealth comes from outpacing the printer. That’s the #Bitcoin journey." 👏

Sam Altman says "eventually, the cost of AI will converge to the cost of energy." Robots can build chips, optimize networks, "but an electron is an electron." In the end, intelligence will scale as far as the grid allows.

@flavioAd o3 and o4-mini the ball pass through each other, that’s not realistic. Shouldn’t the balls bounce of one another like it does on Gemini 2.5 Pro ?

OpenAI released o3 and o4-mini! I've tested both against the top reasoning models using the exact same prompt

@AndrewCurran_ I think we at or near AGI, when it can play common games like Chess successfully without hallucinating like crazy conjuring Queens, Rooks, Bishops and Knights from nowhere.

Tyler Cowen has seen enough.

@DeryaTR_ @OpenAI Hello Sir… can it play Chess properly without hallucinations?

I’m absolutely blown away by @OpenAI’s new o3 model! I’ve had early access and haven’t put it down for days. This release feels like the milestone we experienced with o1-preview and o1-pro, but smarter and more reliable in every way, it truly cranks everything up to eleven! In fact, I feel o3’s intelligence is at or near genius level! It never hallucinates, and its new agent-style tools effortlessly handle multi-step tasks with incredible reasoning and precision, generating complex incredibly insightful and based scientific hypotheses on demand! When I throw challenging clinical or medical questions at o3, its responses sound like they’re coming directly from a top subspecialist physicians: precise, thorough, confidently evidence-based, and remarkably professional, exactly what you’d expect from a genuine expert on the topic. I feel it is also clearly superior to the current SOTA model Gemini 2.5-pro in these aspects, and more intelligent. Without reservation, I confidently claim this is a total game-changer for science, medicine, and likely many other fields! Kudos to the entire amazing team at OpenAI, there’s no doubt this advance will continue pushing humanity forward and significantly improve our lives! 👏👏👏

@ChrisCamillo I see the vision and align with this message

Long-term AI + embodied AI productivity boom > short-term tariff, trade war, and rate noise.

@togethercompute @Agentica_ How does it compare to QwQ that is also a 32B parameter reasoning model?

Announcing DeepCoder-14B – an o1 & o3-mini level coding reasoning model fully open-sourced! We’re releasing everything: dataset, code, and training recipe.🔥 Built in collaboration with the @Agentica_ team. See how we created it. 🧵

@TalvenArt I think Neva just healed or restored energy. This would line up with what Goku said about already having SSJ4. Else this is a retcon or plot hole on what exactly Neva did regarding SSJ4🤦‍♂️

I feel like people are misinterpreting what was being said here. I don't it's a case of Goku having ssj4 the whole time, and more of he knew there was something there but wasn't able to reach it until Neva did whatever the frick he did. But that's just how I feel.

@ChrisCamillo LPDDR5X 👀

Back to work. #socialarb

@ErikVoorhees So they have a copy of my sh*tty random prompts and spaghetti code. Awesome I’ve done my part to help them. I look forward to DeepSeek-R2.

When you use DeepSeek app, all your convo, IP, prompts, etc are going to Chinese government. When you use DeepSeek through the Venice.ai app, none of it is going anywhere. DeepSeek is for Pro users only, and of course you can pay in crypto, no email required even.

@AJStavole @unusual_whales TikTok would be more. You give it access to things like camera and file/folders. In addition to your digital footprint.

@unusual_whales So it’s Tik Tok on steroids

DeepSeek reportedly collects IP, keystroke patterns, device info, and stores it in China, per TC

@unusual_whales Just like every online company. It would be weird if they didn’t tbh.

@JonathanHeindl @ThePrimeagen Those are not Deepseek they are distillation. Completely different from Deepseek tbh. (Basically fine tune of Llama or Qwen depending on param size)

@ThePrimeagen there are smaller versions 🤔 so u can try just running it on ur current machine , also i heard its worth an attempt to run it in ram rather than on gpu

Give me a list of things I need to buy to use r1 in all it's glory If I could just have a motherboard with a couple gpus plugged in, thrown together with hot glue and PVC pipe that would make me a happy man

@JeremyCMorgan @ollama @deepseek_ai Do you find it typically taking around 1min 16sec to complete prompts?

@ollama @deepseek_ai Wow. 32b model, running 37 tokens/s on an RTX 4090. Impressive!

DeepSeek's first-generation reasoning models are achieving performance comparable to OpenAI's o1 across math, code, and reasoning tasks! Give it a try! 👇 7B distilled: ollama run deepseek-r1:7b More distilled sizes are available. 🧵