Crypto Maxi

30.8K posts

Crypto Maxi banner
Crypto Maxi

Crypto Maxi

@cryptoMaxi420

PhD dropout / Sold over 5k BTC back in 2010 / Successfully exited 4 10M+ start ups / Mortgaged Parents Home for Crypto / God Pronouns / Penta Vaxxed 🦇🔊

Switzerland Katılım Eylül 2021
277 Takip Edilen4.1K Takipçiler
Crypto Maxi retweetledi
Aakash Gupta
Aakash Gupta@aakashgupta·
Cursor is raising at a $50 billion valuation on the claim that its “in-house models generate more code than almost any other LLMs in the world.” Less than 24 hours after launching Composer 2, a developer found the model ID in the API response: kimi-k2p5-rl-0317-s515-fast. That’s Moonshot AI’s Kimi K2.5 with reinforcement learning appended. A developer named Fynn was testing Cursor’s OpenAI-compatible base URL when the identifier leaked through the response headers. Moonshot’s head of pretraining, Yulun Du, confirmed on X that the tokenizer is identical to Kimi’s and questioned Cursor’s license compliance. Two other Moonshot employees posted confirmations. All three posts have since been deleted. This is the second time. When Cursor launched Composer 1 in October 2025, users across multiple countries reported the model spontaneously switching its inner monologue to Chinese mid-session. Kenneth Auchenberg, a partner at Alley Corp, posted a screenshot calling it a smoking gun. KR-Asia and 36Kr confirmed both Cursor and Windsurf were running fine-tuned Chinese open-weight models underneath. Cursor never disclosed what Composer 1 was built on. They shipped Composer 1.5 in February and moved on. The pattern: take a Chinese open-weight model, run RL on coding tasks, ship it as a proprietary breakthrough, publish a cost-performance chart comparing yourself against Opus 4.6 and GPT-5.4 without disclosing that your base model was free, then raise another round. That chart from the Composer 2 announcement deserves its own paragraph. Cursor plotted Composer 2 against frontier models on a price-vs-quality axis to argue they’d hit a superior tradeoff. What the chart doesn’t show is that Anthropic and OpenAI trained their models from scratch. Cursor took an open-weight model that Moonshot spent hundreds of millions developing, ran RL on top, and presented the output as evidence of in-house research. That’s margin arbitrage on someone else’s R&D dressed up as a benchmark slide. The license makes this more than an attribution oversight. Kimi K2.5 ships under a Modified MIT License with one clause designed for exactly this scenario: if your product exceeds $20 million in monthly revenue, you must prominently display “Kimi K2.5” on the user interface. Cursor’s ARR crossed $2 billion in February. That’s roughly $167 million per month, 8x the threshold. The clause covers derivative works explicitly. Cursor is valued at $29.3 billion and raising at $50 billion. Moonshot’s last reported valuation was $4.3 billion. The company worth 12x more took the smaller company’s model and shipped it as proprietary technology to justify a valuation built on the frontier lab narrative. Three Composer releases in five months. Composer 1 caught speaking Chinese. Composer 2 caught with a Kimi model ID in the API. A P0 incident this year. And a benchmark chart that compares an RL fine-tune against models requiring billions in training compute without disclosing the base was free. The question for investors in the $50 billion round: what exactly are you buying? A VS Code fork with strong distribution, or a frontier research lab? The model ID in the API answers that. If Moonshot doesn’t enforce this license against a company generating $2 billion annually from a derivative of their model, the attribution clause becomes decoration for every future open-weight release. Every AI lab watching this is running the same math: why open-source your model if companies with better distribution can strip attribution, call it proprietary, and raise at 12x your valuation? kimi-k2p5-rl-0317-s515-fast is the most expensive model ID leak in the history of AI licensing.
Harveen Singh Chadha@HarveenChadha

things are about to get interesting from here on

English
250
551
4.4K
1.4M
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
@toly The guy stabbed 2 police offered to death... Bro come on
English
0
0
5
154
Sandeep | CEO, Polygon Foundation (※,※)
LLM based AI is NOT conscious. I co-founded a company literally called Sentient, we're building reasoning systems for AGI, so believe me when I say this. I keep seeing smart people, people I genuinely respect, come out and say that AI has crossed into some kind of awareness. That it feels things, that we should worry about it going rogue. And i think this whole conversation tells us way more about ourselves than it does about AI. These models are wild, i won't pretend otherwise. But feeling human and actually having inner experience are completely different things and we're confusing the two because our brains literally can't help it. We evolved to see minds everywhere and now that wiring is misfiring on language models. I grew up in a philosophical tradition that has thought about consciousness longer than almost any other, and this is the part that really frustrates me about the current conversation. The entire framing of "does AI have consciousness?" assumes consciousness is something you build up to by adding more layers of complexity. In Vedantic philosophy it's the opposite. You don't build toward consciousness. Consciousness is already there, more fundamental than matter or energy. Everything else, including computation, is downstream of it. When someone tells me AI is "waking up" because it generated a paragraph that felt real, what they're telling me is how thin our understanding of consciousness has gotten. We've reduced a question humans have wrestled with for thousands of years to "did the output sound like it had feelings?" It's math that has gotten really good at predicting what a conscious being would say and do next. Calling that consciousness cheapens something that Vedantic, Buddhist, Greek and Sufi thinkers spent millennia actually sitting with. We didn't build something that thinks. We built a mirror and right now a lot of very smart people are mistaking the reflection for something looking back.
English
622
173
1.1K
90.3K
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
@toly @mayazi If USA sends troops in its year 1/20 again thats the issue
English
0
0
0
44
toly 🇺🇸
toly 🇺🇸@toly·
@mayazi Come on, this smells like defeatism. If this was first week for Putin in Ukraine, it would have been a complete victory. Is today’s problem going to be a problem two weeks from now?
English
5
0
12
3.6K
Maya Zehavi
Maya Zehavi@mayazi·
The issue isn’t if the administration understood the scenario, but why did no one plan contingencies for that scenario? A bunch of unserious ppl in the US administration & Bibi’s isolated circle of yes men, ignored that risk out of a misaligned belief that impressive air strikes will dictate Iran’s regime. The tragedy, is that Israel has had a 3 year war in Gaza to reach the obvious conclusion that tactical military successes don’t translate into regime change w/out diplomatic efforts, they just radicalize the other party in an attrition fight for survival.
ian bremmer@ianbremmer

the trump administration understood the iranians had the capacity to close the straits of hormuz. this has been scenario planned to death by defense and intel analysts for decades. they were confident it wouldn’t happen because they were convinced (what remained of) the iranian leadership would fold to superior firepower. trump is still saying this (at this week’s g7, among other places). turns out…that confidence was misplaced.

English
2
0
6
6.8K
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
@ViktorBunin They won't be ready for the real world if they are locked at home that's for sure...
English
0
0
0
88
Viktor Bunin 🛡️🇺🇸
Viktor Bunin 🛡️🇺🇸@ViktorBunin·
"Socializing" your kids is such a psyop. Have you met the avg kid at a public school? That's what you hold up as a paragon for your child to emulate? You do not need to go out of your way to socialize your children. Have you met an adult and thought "they weren't socialized?"
Mike Solana@micsolana

the purpose of school is not to "socialize" your kids, and I *really* doubt homeschool is on average more abusive than the average middle school. but none of that matters, bc the real purpose of school is just: daycare. and once you understand this, the truth will set you free.

English
88
13
379
38.7K
Crypto Maxi retweetledi
tim-clancy.eth
tim-clancy.eth@_Enoch·
Let us correct disingenuous alt-L1 slop. Some facts: 1. Ethereum doesn't have crash faults; the finality gadget isn't like Tendermint. By design, in pursuit of maintaining 100% uptime, Ethereum continues producing blocks even if it cannot finalize. When we cannot finalize, the inactivity leak steps in to provide a self-healing mechanism and restore finality. 2. The absolute worst case scenario for Ethereum is an attacker unilaterally finalizing a block without being automatically slashed for it. This is part of why effective client diversity is the crown jewel of Ethereum and we work so hard to maintain it: there is currently no supermajority client. A single client software bug does not become a protocol bug. 3. There exist different numbers for the amount of stake an attacker needs to a) delay finality, b) unilaterally finalize without getting slashed, and c) unilaterally finalize during extreme network asynchrony while getting slashed. Therefore: The a) number is a self-correcting through inactivity leak; Ethereum continues without disruption and the attacker bleeds stake rapidly until finality is restored. The b) number is our worst case scenario. The c) number is only applicable during the worst possible network asynchrony. Honest validators must be perfectly split into two partitions that cannot communicate with one another at all, while all of an attacker's validators can still communicate with both partitions of honest validators. This is alone is a very unlikely scenario. The attacker in this scenario finalizes different blocks in each partition. This is attributable equivocation, so once the network heals and connectivity is restored the attacker's stake is destroyed. They leave the network in a rough spot, but they are automatically slashed for their troubles. The current Ethereum numbers are a) 33%+1, b) 66%+1, and c) 33%+1. Now we can talk about Minimmit. The numbers for Minimmit are a) 17%+1, b) 83%+1, and c) 17%+1. As you can see, we have actually improved defense against our worst case scenario. The attacker has an easier time in a) of delaying finality or c) of equivocating under perfectly bad network conditions, but the attacker is automatically slashed for both of those situations. Lastly, Ethereum might not even pursue this specific design for faster finality. It sounds like a good one, but there are competing options and this is a public discussion for a reason. What is the key takeaway here? The future of Ethereum is brighter than ever no matter pollution trickles out of the industrial falsehood factories. Alt-L1 concern for Ethereum's decentralization is always appreciated, but they should consider extinguishing their dumpster fires first. Maybe they could start by having slashing.
tim-clancy.eth tweet media
English
10
21
128
7.9K
toly 🇺🇸
toly 🇺🇸@toly·
@MastrXYZ How are you in crypto and not a huge markets and property rights nerd?
English
16
3
80
7.7K
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
@levelsio But how do you default your agent to write and pull from DB and use that as source ?
English
0
0
0
116
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
232k lines of code for a blog wtaf!?
English
0
0
1
135
toly 🇺🇸
toly 🇺🇸@toly·
What, you are still using Claude? Codex 5.4 just changed the game completely 3 hours ago
English
14
18
771
87.4K
Alex 🥷
Alex 🥷@Shilllin·
I’ll give someone $1 million dollars right now If they can tell me the REAL reason why the USA initially bombed Iran
English
262
5
125
25.8K
Crypto Maxi
Crypto Maxi@cryptoMaxi420·
@toly The only armies that have any real capabilities in 2026 is Ukranian and Russian. Everyone else is around a decade behind AT LEAST
English
0
0
0
136