Shin Kato

20.4K posts

Shin Kato banner
Shin Kato

Shin Kato

@charuman

Product at @Amazon AGI. Multilingual LLM Previously at @creditkarma, @Cruise, @awscloud, @amazon. Love building products for builders. Morioka city :-)

Berkeley, CA Katılım Haziran 2007
1.2K Takip Edilen1.7K Takipçiler
Shin Kato
Shin Kato@charuman·
I love people who stick with their moral compass and own principles, be they good or bad.
English
0
0
0
54
Shin Kato retweetledi
Arun Rao
Arun Rao@sudoraohacker·
OpenAI was struggling when it lacked focus, had silly side projects, and had excessive DC agreements for scaling (putting it at great financial risk). Cutting side projects to focus on reasoning/agentic coding and work models is the right move, as is trimming DC spend. The big open question is whether they still want to win the consumer market vs Google and Meta, and for that they need an ad business that works.
i/o@avidseries

My impression is that OpenAI seems to be struggling on a number of different fronts. Anybody else seeing this?

English
0
1
6
966
Shin Kato retweetledi
Yuta Kashino
Yuta Kashino@yutakashino·
日本では話題になってないけど,litellm 1.82.7, 1.82.8 がサプライチェーンアタックによってマルウェアが仕込まれてクレデンシャルを盗まれる.litellmをデフォルトで利用するGoogle ADK, DSPy, CrewAI, smolagents, Microsoft GraphRAGを最近インストールした人はチェックしたほうがいい.
Yuta Kashino@yutakashino

Supply Chain Attack in litellm 1.82.8 on PyPI futuresearch.ai/blog/litellm-p… オオウ…。今はコーディングエージェントがあるから、スキルがなくてもこういう下位ライブラリにマルウェア仕込んでサプライチェーン攻撃を行うことができるからな…

日本語
1
7
25
6.1K
Shin Kato retweetledi
Aakash Gupta
Aakash Gupta@aakashgupta·
OpenAI just exited the video generation business entirely. App dead. API dead. No video inside ChatGPT. Disney’s $1 billion deal, signed four months ago, is dead. Read that again. This isn’t a consolidation into the super app. Altman told staff Tuesday that OpenAI is winding down all products using video models. Disney’s own statement says they respect OpenAI’s decision to “exit the video generation business.” The Sora research team is being redirected to robotics. The reason is sitting right there in the competitive data. Anthropic hit $19 billion in annualized revenue by early 2026 selling text and code. No video generation. No image generation. No consumer social app. No Disney deal. One product surface: chat, code, computer use, all in one place. OpenAI looked at where every dollar of market growth was coming from and saw the answer: coding and enterprise. So now they’re copying the model. ChatGPT, Codex, and the browser merge into one app. Instant Checkout killed today too. Every consumer experiment is getting cut. What remains is the Anthropic playbook: one app, code and chat, enterprise and developer focus. The Sora numbers explain the urgency. Total consumer revenue across iOS and Android since September: $1.4 million. Peak month was $540,000. Every video generation burned GPU compute that could have been running inference for ChatGPT or Codex instead. OpenAI’s own head of Sora announced generation limits because chips couldn’t keep up. At $14 billion in projected 2026 losses, every GPU matters. Google just inherited the AI video market by default. Nano Banana already lives inside Gemini. No standalone app to manage, no separate brand to support. Among the majors, they’re the only ones left. Runway, Kling, Minimax, Luma, and the other independents are still shipping, but none of them have Google’s distribution. Disney put $1 billion in stock warrants on a product that lasted six months. The deal was announced in December. Characters from Marvel, Pixar, and Star Wars were supposed to be generating fan videos on Sora by now. Instead, Disney is writing a polite press statement about “respecting OpenAI’s decision” while its legal team unwinds a deal that never produced a single licensed video. Four months from billion-dollar partnership to obituary. That’s how fast the AI product landscape reprices when the unit economics don’t work.
Sora@soraofficialapp

We’re saying goodbye to the Sora app. To everyone who created with Sora, shared it, and built community around it: thank you. What you made with Sora mattered, and we know this news is disappointing. We’ll share more soon, including timelines for the app and API and details on preserving your work. – The Sora Team

English
233
961
7.7K
1.8M
Shin Kato
Shin Kato@charuman·
Why do people want Mac Mini where much cheaper mini PCs can host OpenClaw?
English
0
0
0
118
Shin Kato retweetledi
Rie Yano やのりえ
Congrats on the launch @yusan_lin and @mirrormirror_ai!
Yusan Lin@yusan_lin

Today @mirrormirror_ai is launching the marketplace where fashion models license their likeness and brands get stunning AI-generated imagery featuring real people. Commercially licensed, model-approved. Try our platform: mirrormirrorai.com As a fashion model I used to spend hours on fashion photoshoot sets. I later did my PhD in CS and became a Research Scientist on AI for fashion. I can see clearly that AI image generation is replacing a large portion of my old job. But brands that use AI recklessly have already paid the price. It damages reputations and hurts the bottom line. Putting real people at the core of AI-generated imagery isn't just about avoiding backlash. It's better business. That's what Mirror Mirror AI is built for. Right now, Mirror Mirror AI houses agency-signed models who have graced the covers of Vogue and Harper's Bazaar. You can digitally book them using our fashion-centric AI software, get your campaign done in hours instead of weeks, and never have to fly anyone in. You purchase a license for commercial use upon approval, and the models get paid. Mirror Mirror AI is also opening a global call for independent models from anywhere in the world to apply to be featured on the platform. Work with fashion brands internationally, choose the projects you take on, and earn from your own likeness on your own terms. Selected models will be announced at an exclusive event in New York during @Techweek_ this June. Apply for the open call: mirrormirrorai.com/open-call A huge thank you to our incredible team for pouring their hearts into this launch, and to a16z @speedrun for believing in our vision from the start. We're just getting started.

English
0
1
5
740
Shin Kato retweetledi
Arun Rao
Arun Rao@sudoraohacker·
Some of the Olmo / Ai2 team just got poached by Microsoft. I don’t think the small labs will be able to keep talent, as the big ones compete fiercely and the stakes keep getting higher on model building (MSFT needs more in house models and a major Copilot refresh). Curious to see if @natolambert sticks around or leaves for Nvidia or someplace else.
Arun Rao tweet media
English
0
2
6
988
Shin Kato
Shin Kato@charuman·
OpenClaw、物忘れが激しすぎる。
日本語
0
0
0
125
Shin Kato
Shin Kato@charuman·
WTIとブレントは相関するんだから、ホルムズ海峡が封鎖され続けたら、原油自給率が高いとか関係ない。共和党みんな焦ってると思う。中間選挙近いし。
カズ in シアトル@kaz_kazuhisa

原油価格高騰にアメリカ側が焦ってる印象。長引かせたくないから「発電所攻撃」をちらつかせて脅したけど効かなかった。もしお互いにインフラ攻撃を始めてしまうと、ますます落とし所が無くなるように思う。

日本語
0
0
1
422
Shin Kato
Shin Kato@charuman·
ある程度優秀なベースモデルを特定機能特化型のモデルにできるかというのは割とアプリレイヤーの会社の競争力を左右するのではないかと思う。4Bから12BみたいなサイズのモデルにCPT、事後学習で汎用プレミアムモデル並みの性能を引き出している会社もいる。
日本語
0
0
1
150
Shin Kato retweetledi
Masahiro Tanaka
Masahiro Tanaka@toh_tana·
ピッツバーグに着いた! ASPLOSで共著論文が採択されたのと、チュートリアルのプロポーザルが通って講演があるので、現地で参加。 一月のラスベガスに引き続き、今年2回目のカジノ。CESはまあ分かるけど、学会はカジノなくても良くない?タバコの匂いがきつい。
Masahiro Tanaka tweet media
日本語
0
1
9
797
Shin Kato
Shin Kato@charuman·
8月から続いていた戦いがひと段落。まだ続くけれど一番きついところは乗り越えた。計り知れないほどのサポートをもらい感謝しても仕切れないし、色々あるけどやっぱりアメリカいい国だなってなった。 ちょっと燃え尽きてる。温泉行きたい。
日本語
0
0
5
280
Shin Kato
Shin Kato@charuman·
まあしかもいくつかチェックポイント用意してベンチマークのバランスを見ながら、SFTからRL(GRPO/DPOなどに)どれを移すか決めたりするし、YOLOみたいな場合もあるだろうけど稀では。
Yuki Nakazato@ynakazat

米国のでかいテック企業でも、LLMの学習は長期に亘るので、学習中に途中のチェックポイントからデータのミックスを変えたりとかlearning rate変えたりとかというのは当たり前のように行われるし、これを知らない人がその解像度で国産LLMに反対しているんだったらちょっとウケるな

日本語
0
0
3
468
Shin Kato
Shin Kato@charuman·
That’s what I felt. Nvidia is now an ecosystem developer. Anything that can thrive in the ecosystem is more than welcome.
Aakash Gupta@aakashgupta

Jensen Huang dedicated a full segment of his GTC keynote to an open-source project he didn’t build, doesn’t own, and can’t monetize directly. Then he called it “the most popular open-source project in the history of humanity.” This tells you everything about what NVIDIA is actually selling. OpenClaw is an AI agent that runs locally on your machine, 24/7. It needs a computer that stays on around the clock. When the project went viral in January, Mac Minis sold out overnight. People were buying 40 at a time for batch deployment. Alibaba Cloud started running a $9.90/month server promotion specifically for OpenClaw users. DigitalOcean launched one-click deployment. Every single one of those machines needs compute. Every business scaling agents needs GPUs. NVIDIA moved faster on OpenClaw than they’ve moved on anything since CUDA. They built NemoClaw, a full enterprise security and privacy stack, and gave it away for free. NVIDIA doesn’t do free. They do free when the downstream hardware pull justifies the investment ten times over. At the same keynote where Jensen compared OpenClaw to Windows, Linux, and HTML in the same sentence, he announced Blackwell and Vera Rubin purchase orders tracking toward $1 trillion through 2027. Up from the $500 billion projection last year. The framing is genius. Jensen positions NVIDIA as the enabler of an open-source revolution built by the community. “Every carpenter can now be an architect. Every plumber will become an architect.” The crowd cheers. Meanwhile NVIDIA is the only company selling the shovels, the picks, and the mine itself. DGX Spark. DGX Station. Vera Rubin racks. All purpose-built for always-on agent workloads that didn’t exist six months ago. 250,000 GitHub stars. One developer who built the prototype in an hour and left for OpenAI a month later. Peter Steinberger’s own explanation for why the big labs didn’t build it first: “It’s not a technical issue but an organizational-structure problem.” And now the CEO of the world’s most valuable company is on stage doing the marketing for free, because the “operating system for personal AI” is also the greatest GPU demand generation engine anyone has ever built for him, and he didn’t have to spend a dollar creating it. Jensen told the GTC audience the largest percentage of attendees were from financial services. Then joked, “I’m hoping it’s developers, not traders.” They both heard exactly what they needed to hear.

English
0
0
0
227