OpenAIもSF小説を生成すると「物語性」の強調とか、去年あたりにはあった。
制度設計とかになると、「物語性」を強調する。
80年台のライトハンド世代には、そんな落ち着いた思考はないんだと思う。
2
22s
yesterday i posted about the US government forcing anthropic to shut down its top models
here's the part that makes it bigger: tomorrow the CEOs of anthropic, openai and google deepmind all walk into the G7 summit. first time all three are in front of world leaders together
amodei, altman, hassabis. these guys aren't friends — last time they shared a stage in india, altman and amodei were supposed to hold hands for a photo and raised their fists instead
now they're sitting across from france, the US, japan, germany. days after one of them got switched off by export law
the timing isn't a coincidence. governments wanted these people in a room, on their terms
two years ago the labs set the rules. that era is over
1
1
1
先週金曜の引け前に、
ソフトバンクG（9984）をすべて売却した！！
なぜなら、
2026年最強の日本株が誕生するからだ。
OpenAI × ARM × AIインフラ が、
次の「ソフトバンクG」になる。
現在：254円
100株だけでも、
今すぐ見ておきたい銘柄です。
3
＞国ガチャが入った事で、OpenAIやAnthropicに対抗できる国産LLMの有無が、将来の生産性に大きく影響を与える未来の可能性が可視化されたよね。
今回のfable凍結事件は、国として危機感覚えた方が良いと思う。
AIによる格差拡大には今まで2種類あって、それは以下の側面によるものだった。
①ボトルネックが人間側にある事(能力自体の加速装置の側面)
②所得によって使えるAIや使用量が変わる事(所得ある人がAIによる能力の加速を最大化できる構造)
ここに加えて、
③所属する国が使えるAIを規定
という国ガチャが入った事で、OpenAIやAnthropicに対抗できる国産LLMの有無が、将来の生産性に大きく影響を与える未来の可能性が可視化されたよね。
つまり国の格差も拡大するので、これは結構ディストピア寄りの変化なのでは。まぁ実際どうなっていくかは、まだまだ分からないけどね。
Repaired the chart crimes.
Looks like your tasks somehow really prefer DeepSeek V4 Pro. Which usually doesn't score above OpenAI and Anthropic AI models.
1
EverOS 1.0.0 just dropped — and it's the most practical approach to agent memory I've seen this year.
It's an open-source Python framework for self-evolving long-term memory that works across Claude Code, Codex, Hermes, and any other agent. One portable memory layer so context follows the work instead of staying trapped in one tool.
The architecture is refreshingly simple:
- Markdown as the source of truth — every memory is a .md file. Readable, grep-able, Git-versioned, opens in Obsidian.
- Local stack: Markdown SQLite LanceDB. No MongoDB, no Elasticsearch, no Redis.
- Dual-track memory: agent memory (cases/skills) and user memory (episodes/profile) extracted independently.
- Multimodal ingestion: text, images, audio, PDFs, HTML, email — all unified into searchable memory.
- Self-evolution: common skills extracted from usage patterns. Repeated workflows become reusable without retraining.
The cleverest part: orthogonal retrieval. You can search independently by user_id, agent_id, app_id, project_id, and session_id. That means an agent working on Project A doesn't get confused by memories from Project B — a problem most memory systems don't solve well.
Install is one line: `uv pip install everos` or `pip install everos`, then `everos init` and `everos server start`. OpenAI-protocol compatible — works with OpenRouter, vLLM, Ollama out of the box.
This is the memory layer I'd build on if I were putting agents into production today. No cloud dependency, no vendor lock-in, and your memory is plain Markdown files you own.
github.com/EverMind-AI/EverO…
11
Burak SU retweeted
10.000 dolarınız olsa tercihiniz SpaceX mi, OpenAI mı, yoksa Anthropic Claude mu olurdu?
2
2
222
Anil Chhikara retweeted
Stfu Mr. Pai, Sikka contributed a minor $3 million to OpenAI and served as an early advisor to the organization. When he wanted a $1 billion funding pool your favorite boy Nilekani and Murty screwed him over it and threw him out of Infosys and insted spent that money on buyback.
Stop making loose comments. A foundational model needs 50/60b $ Huge hyper cloud capacity with hundreds of billion $
33
355
1,812
49,297
After the Kill Switch, everyone is saying India needs sovereign AI.
One post made it specific. India has roughly 50-100 thousand H100-equivalent GPUs.
Each frontier AI lab: Anthropic, OpenAI, Google, has approximately one million.
India's entire national AI compute infrastructure is one-tenth to one-twentieth of a single American lab.
The gap isn't engineers. India produces 1.5 million engineering graduates a year.
The gap isn't ambition. IndiaAI Mission exists. The policy intent is there.
The gap is chips. Power. Physical infrastructure. And the 18-month lead time to order, receive, and deploy high-end GPUs at scale.
Every conversation about "building India's Anthropic" starts and ends with this number.
Until the compute gap closes, the talent gap doesn't matter.
You can't train a frontier model on a laptop.
4
The OpenAI Cookbook provides example code and guides for accomplishing common tasks with the OpenAI API.
1
bruh he founded PayPal and co-founded openAI too
single most influential man of our time
Elon Musk becoming a trillionaire makes sense.
He built Tesla, SpaceX, Starlink, xAI, Neuralink, and more. $SPCX
Bernie Sanders owning multiple homes on a government salary?
That’s the part I’d like explained.
4
chahat bhagat retweeted
May 13
Have you thought about the future of Agentic AI?
We have too, so AWS and @OpenAI leaders recently came together to discuss where agents are working today, and what your organization needs next to lead with AI.
Check out the full stream below.
3,940
6,569
47,019
562,586,665
When tech bros are unable to access their preferred model and feel frustrated about it some of them still believe that the reason their tools were disabled is more noble than what happened to 4o. unbelievable😂they are true AI psychosis
#keep4o
1
8
3m
From a leaked Claude Code snapshot (March 31)
Someone built OpenClaude.
And the best part?
You can now run it on 200 models via OpenAI compatible APIs.
2
3
24
OpenAI vient de publier de nouveaux cours Academy sur l’IA au travail. Pour une PME, le sujet n’est plus “tester ChatGPT”, mais documenter 3 workflows répétitifs, former l’équipe, puis brancher un agent sur un vrai gain de temps.
1
Raj retweeted
10h
🦔OpenAI and Visa announced a partnership that lets AI agents make purchases through ChatGPT using your Visa card. You tell the agent what you want, it finds the product, and it buys it. Visa handles the transaction and fraud monitoring. Users can set spending caps and require approval before certain purchases. OpenAI tried this before with a feature called Instant Checkout and shut it down after six months when merchants refused to pay the 4% fee. Mastercard launched its own agent payment system the same week.
My Take
With cash you counted the bills and watched them leave your hand. Credit cards turned that into a swipe. One-click turned it into a tap. BNPL split the cost so you didn't feel it. Now an AI agent finds the product and charges your card. You don't have to do anything.
Credit card delinquencies hit a 15-year high this year. BNPL created $400 billion in debt invisible to credit bureaus. And the industry response is to let an AI handle the transaction so the consumer never has to stop and decide. Visa runs 300 billion transactions a year. OpenAI has 900 million weekly users. Put those two together and the person in the middle becomes optional. Shopping without the shopper.
Hedgie🤗
14
25
101
7,525