Assessing businesses for investment https://nitter.gq/t.co/Qc02kllZV7 Tweets are not investment advice.
Joined January 2024
- Tweets5.1K
- Following542
- Followers556
- Likes13.9K
Pinned Tweet
Deepseek the value destroyer
... for hyperscalers wanting to see ROI, Deepseek has reduced the numerator by 95+%, greatly reducing the odds of positive returns
Open-source / open-weight also breaks monopoly pricing power
But for consumers?
The value has shifted to consumers
Let me say the quiet part out loud:
AI model building is a money trap.
What you are seeing here is a feature of modern AI models: there is no bounding law like Moore’s law where advances can be predictably expected. Nor has any model builder adequately publicly explained what to expect sans jargon and BS.
What this means is that it’s very difficult to plan spending programs to create advances and will likely render entire months and possibly years of energy, hardware, R&D, OpEx and CapEx wasted and useless as new innovations are discovered.
Open Source is the clear winner.
Closed source will be forced to keep their best models secret and sell to enterprises OR try and create some incredible consumer app with it.
Having done this before, my lived experience is that this is easier said than done.
Net/net consumers will win, many venture funded efforts will not.
Even a quantized 8-bit 27B local model can beat Opus 4.8?
Would most tasks (or most folks) require a model stronger than Opus 4.8?
What times we live in
If you're hiring for the avg office role
You'll not pay $300k as you might for an ML PhD
Your budget might be $60k
And you'll look for someone w/ their wits about them, but also reliable & WITHIN budget
Human work is like this, so work for AI will also likely be like this
たった今、世界のAI覇権のバランスは完全にぶっ壊れました。これほんまにやばいです。
週明けのAI関連株は大暴落どころの騒ぎじゃないかもな…。
何が衝撃的なのか、ガチ解説します。
もう何もかもが変わっちまったんだな...。
■何がおきたか
DeepSeek v4.1-Flash 1017というモデルが公開されました。世界のAI情勢を破壊します。ヤバいモデルです。
このモデル簡単がどういうものなのか説明すると、
フロンティアモデルを一部で越える（SOTA級）のモデルを中華ラボがいきなり出してきた。
問題はその性能、値段とサイズと仕組み
どれも本当に破壊的なものだから、1個ずついきます。
■性能
全分野でOpus5.5を超えたっていうわけじゃないけど。
完全にDeepSeek v4.1-Flash 1017はSOTA級
DeepSWEっていうベンチが一番衝撃的かな。
- 今回のFlash：成功率77%、約0.20ドル／タスク
- GPT-6.1 Solの最も高得点の設定：約75%、約0.65ドル／タスク
つまり、この比較では約3分の1の費用で、成功率も少し高い。大量に実装・修正を回す用途には大きな差。
というか、成功率でSolを越えてるっていう事実がヤバいんだよ。
■値段
もう完全に乱暴に言うと。ほぼタダです。<言い過ぎ
INPUT$0.15Output$0.3
入力10万＋出力1万トークンの例だと
Solより約86〜93%安く、Opusより約93〜97%安い
DeepSeekハーネスを使えば、自分の利用では、入力のキャッシュヒット率が99%になることもあったので。
実際はもっと安い。
過去に湯水のように使えるAstra,Opus,Fableをくれとは言ったが。ここまでやれとは言ってない。
本当に、なんだろう、マジで安いんだよ。
実際の感覚だと200倍ぐらい安いね。
■サイズ
これだけヤバいモデルのオープンウェイトが同時に無料公開されました。
そのサイズは763B。重みだけで約510.3GB
例えば米国のフロンティアモデルのサイズは10Tとか言われてるんですね、公開情報じゃないからわからん無いけど。Kimi K3とかが2.8T
要するに、めちゃくちゃサイズが小さいんですよ。
このパラメータ数って多ければ多いほど計算に計算資源と電力がかかるんで、コストかかるんですよ。
アメリカのAIモデルって、バカみたいに計算資源を保有してるからゴリ押しでそれを解決してるんですけど。
こんだけ軽かったら、もう普通に家で使えるレベルなんですよ。
実際、DGXSpark × 4台とかMacStudio 512GBで動かせると思います。
バカみたいなデータセンターがなくても、フロンティアモデルが動くようになっちゃったってわけ。
■仕組み
ここがかなり驚き度の高いポイントなんですけど。
DeepSeek-V4.1-Flash自体はもともと存在していて。
まあそこそこ人気だったんですけど。本当に安くてそこそこ性能いいから、結構人気あるよね、ぐらいの賢さだったんですよ。
そのもともとあったモデルを自己フィードバックを使うRLで強化した。って公式は言ってて。
まあ詳しい説明は省きますけど。要するに、既に存在していたモデルを後から調整していくようなやり方で、とんでもなく性能をあげた。
驚くのは、既存モデルの後学習で、開発タスクの成功率を上げながら、処理時間や出力量まで減らしたところです。性能を上げたうえで、使う側の費用まで下げてきた。その性能の上り方が半端じゃないんだよ…意味わかんねぇよ…
DeepSeekのラボが今世界で最もAI、優れたAIラボであるということを認めなきゃいけない。
Monetization matters
AI chatbot platform Character. ai filed for Chapter 11 bankruptcy on Monday and announced it had received a $20 million opening offer to license some of its intellectual property from Disney. However, Disney previously accused the the AI company of unauthorized use of its copyrighted characters.
Read more: forbes.com/sites/siladityara…
📸: Algi Febri Sugita/SOPA Images/LightRocket via Getty Images
Alpha Exponent retweeted
CoreWeave’s unsecured notes are officially 1,000bps STW, now yielding ~15%.
Either this is terribly mispriced and a buy or the HY market is telling you something about the AI build-out
Which one is it?
robinhood:0x5f10a1c971b69e47e059e1dc91901b59b3fb49c3
iran strike options sound like a fancy class of exotics where strike K ∈ {K_high, K_low} is not known until the event happens & can still change ...
Alpha Exponent retweeted
It finally happened
America has entered the frontier open weights model competition. We aren't rolling over to China
Reflection announced Beam, a 501B parameter open weights model that is comparable to GLM 5.2, Qwen 3.8, and Opus 4.8
Yes, Opus 4.8 on your desk
You will need a decent sized machine for this. Probably the Mac Studio 512gb
But if you listened to me back in January when I warned you days like today were going to come, then you're all set
I stand by Opus 4.5 being the most important model release in history. We will now have a model BETTER than that on our desks
If you haven't gotten into local AI yet, it's officially time to do it
Take any computer you have, talk to an AI agent, ask which models you can run
Never been more important to just start
Introducing Beam: a highly efficient agentic open model with 501B total parameters and 23B active.
- Frontier reasoning efficiency
- Advances the Western open frontier on coding & agentic tasks
- Trained end-to-end from scratch
Full weights release this month.
Learn more about Beam: reflection.ai/beam
There's strong temptation for incumbents to benchmaxx - all companies suffer from this, but given the additional internal competition, teams w/in incumbents even more so...
$GOOG
There are two ways AI progress could go very badly and that we must avoid.
First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve people. To ensure that, we need ways to ensure that alignment and safety techniques stay ahead of progress in model capabilities.
Second, we could end up in a world with too much concentration of power. If an extraordinarily powerful AI is used by one person or company to impress their worldview onto everyone else, the results could be extremely dystopian.
Avoiding these two threats requires walking a narrow middle path; for example, one country could gain too much power. Another example is one lab ending up with too much power.
This could be virtually anything, e.g.:
Intentional truth: blunt
Intentional falsehood: diplomatic
Unintentional truth: blabbermouth, Freudian, or both
Unintentional falsehood: careless
I resigned from the Fed today. I spent the last three years devaluing savings at both the Fed & the Treasury. Neither company is acting responsibly. They are racing to hyperinflation, dumping dollars on voters and gambling with our lives. More thoughts below.
Alpha Exponent retweeted
Have been using this model all day yesterday. With its new architecture, this release is just as big as R1 and will redefine all future models. It is such a big difference when you have a model close to 5.6 Sol, but it is 5 times faster and super cheap.
AI overcapacity
Time to ask what's the next major theme
Sam Altman reveals his concern that the industry's compute frenzy is showing its first signs of unsustainable silliness
"I'm not worried about our compute buildout plans. I am worried about the world's compute buildout plans."
"I think we are going to be able to use all of the compute very profitably that we are planning to build."
"But I am seeing the first signs of what feels to me like unsustainable silliness, of random new Neocloud popping up, people claiming that they're going to build gigantic amounts of compute next year that I think they don't have the revenue to support or a buyer."
"I definitely feel some fear about what the world is doing as a whole, although we feel very good about what we've committed to."
Average CN incomes in:
2016 <$10k
2026 >$18k
>80% nominal avg income growth is now a lost decade?
And no, unaffordable & unsustainable price hikes of Big Macs aren't a good thing
But affordable prices are!
Common sense > motivated reasoning AKA "economics"
Alpha Exponent retweeted
Ben Thompson on why we might run out of capital before we run out of compute:
"I'm worried about the timing mismatch in terms of the actual return on investment producing enough revenue to fuel investment.
We're working our way down the capital curve. We started with free cash flow.
The speed with which the tech companies blew through the debt markets is kind of incredible. It took like a year, and now Google's issuing equity.
Nvidia's putting together this $500 billion thing to tap into pension funds and insurance floats and things like that. What's after that? Where does the money come after that?
Ideally we actually flip back to free cash flow funding this, but if there's a gap there, if we don't get there soon enough, then we could have a big blowup.
The railroads had a real duration mismatch. To build a railroad and make money off it was a decade or multiple decades long endeavor, whereas you had to issue money to pay for it in the short term.
And then they ran out of money. That's what happened in the 1870s.
We don't have enough compute because there was insufficient investment made in 2023 and 2024.
So today, when you say there's not enough compute, all the money companies are putting in today doesn't equal compute tomorrow. It all goes into compute in 2029.
I believe in AI. I think it's a real thing. I think the economic impact is going to be astronomical.
You can believe all that and be worried about, are we going to make the bridge to this actually generating the level of returns necessary to continue to fuel this going forward?"
My conversation with @benthompson. Ben has been writing Stratechery for over a decade and remains one of my favorite business thinkers.
We covered a lot. Every important company in the industry and the forces acting on all of them.
- Why he thinks it would be problematic for the US to win the AI race
- Will we run out of money to fund AI
- Google becoming Berkshire Hathaway
- Why ads are amazing
- TSMC, Intel, and Samsung
- Nvidia's invisible price cuts + biggest competitors
- Microsoft, Amazon, Apple, and Meta
I love talking to Ben about everything happening in markets and technology. Enjoy!
TIMESTAMPS
0:00 Intro
0:59 America and the AI Race
8:26 AI’s Funding Problem
15:31 AI’s Capabilities and Limits
20:30 Aggregation Theory, AI, and Ads
31:10 Compute, TSMC, and Intel
47:31 Amazon and Apple’s AI Moats
54:43 The Frontier AI Players
74:07 Nvidia and Commoditized Intelligence
82:52 What Survives an AI Bubble?
