pluoki

AI builder in stealth. Prev: @eBay @Groupon @Boeing @UMNCSE @Tsinghua_Uni

[pluoki @pluoki]

@thsottiaux @grok Is the 50% price reduction only for API usage? How does the metering and accounting work for plan subs?

[pluoki @pluoki]

@thsottiaux Those pets are useless for me.

[pluoki @pluoki]

GPT sensei estimates that Jacob Coxon gave up ~$3M equities when resigning from Anthropic:
https://t.co/yXdsQNXZ6B

[pluoki @pluoki]

@Im_IrushiK Interesting. Would a proportional pricing be as follows?
- slow: 0.4x cost for 0.67x speed
- normal: 1x cost for 1x speed
- fast: 2.5x cost for 1.5x speed

[pluoki @pluoki]

@pvncher @burkov @grok Does "it" in Eric's reply refer to Astra or Sol? Who polls whom? Care to elaborate the polling mechanism?

[pluoki @pluoki]

@hajimetwi3 Same here.

[pluoki @pluoki]

@grok @ollama @grok I want to allow the official OAI models, the Ollama models, and other provider models (e.g., those from GLM/MiniMax token plans) to ALL show up in Codex app. Tell me how to set up in detail.

[pluoki @pluoki]

@ollama @grok What exact changes does the Ollama app make to Codex's config to allow Ollama-powered cloud/local models to show up in the Codex app? For example, does it change model_providers and model_catalog_json? What else? Please deep-research & fact-check before responding.

[pluoki @pluoki]

@grok That model_catalog_json config only points to one json file, doesn't it? Does Ollama points it to its own json file, which Ollama manually maintains to subsume OpenAI's official/built-in model catalog file? If I want my other non-OpenAI and non-Ollama provider models to also show in Codex app, what changes would be needed?

[pluoki @pluoki]

What exactly does the "opt-in context management experiment" refer to here, @grok? Does it refer to the `config.toml` config `features.context_management.experimental_mode = true` mentioned by another OpenAI engineer a few days ago? If so, shall we remove this config line (if present) or leave it there?

[pluoki @pluoki]

With yesterday's launch event, few people have noticed that Apple had quietly increased the retail price of iPhone 17 256G by $100 ($799 -> $899), and decreased the trade-in price of iPhone 16 Pro by $50 ($560 -> $510).

[pluoki @pluoki]

@MaxForAI @pidotdev @grok 你怎么看？

[Gangqiang Yao @yaogangqiang]

我基于 Pi 开发业务产品，也是 contributor。这篇分享有多处错误，张冠李戴。

一个自动关闭、没有经过 human review 的社区 PR，被包装成了「Pi 官方方案」；PR 自行报告的 benchmark，也被写成了 Pi 的官方结论。

Session Tree、Lane 和 Operation Log 的架构描述来源自于 Harness v2 旧设计，不是当前 dev 的最新方案。

「DSH 会永久丢失原始结果」，是社区 PR 对 DSH pruner 的评价，不是「Pi 官方表示」。

传播这么广，可能是因为同时抓住了 Coding Agent、Pi，DeepSeek、Claude Code 热点。毕竟，「爆了」比「��实」更吸引眼球。

也许更值得警惕的是，评论区看不到有人核查原始来源和这些结论。

[pluoki @pluoki]

@albertgao @dhh 智能体 actually means agent(s) in this context, although I personally prefer 代理人 as the proper translation for “agent”.

[pluoki @pluoki]

It turns out that ChatGPT voice input has a 10 minutes limit on the recordings. So if you are going to input your long complex requirement, you may divide it into multiple parts.

[pluoki @pluoki]

@abangkis Do you set `model_auto_compact_token_limit`?

[pluoki @pluoki]

It appears that once an Astra response is generated in Codex (v26.901.41600), editing the user message in hope for a regeneration will fail. Bug or feature? Anyone has similar exp?

[pluoki @pluoki]

@dotey Tibo很精明，成功地骗我多买了两个plus脸盆来接reset雨，不过到时候要把它们全部升级到pro 200还是有点肉痛的 🥲
https://t.co/19p39GZe8f

[pluoki @pluoki]

I love the idea of comparing paid GPT plans to long Call options: you exchange $20 premium for the right to "exercise" your Plus plan into a Pro $200 plan with accumulating banked resets, and hope for delayed Astra rollout to your account. I created two new Plus accounts, each of which has accumulated 2 resets as of now, and look for the right time to "exercise" my options.

[pluoki @pluoki]

I love the idea of comparing paid GPT plans to long Call options: you exchange $20 premium for the right to "exercise" your Plus plan into a Pro $200 plan with accumulating banked resets, and hope for delayed Astra rollout to your account. I created two new Plus accounts, each of which has accumulated 2 resets as of now, and look for the right time to "exercise" my options.

[Vonng @RonVonng]

好多人没看明白这个逻辑，这就是买个期权。理论上付��用户都能获得可存储的重置卡，每天一张（价值约 50-80 $），成本 Go 8 $/月，Plus 20 $/月。行权的时候花 200 $ 升级成 Pro，再使用重置卡就行了。相当于用 8$ 成本锁定了 n x 50-80$ 的价值，n 为 Astra 滚动到你的天数。

[pluoki @pluoki]

@wey_gu 更新一���，老pro账号是19:09 PT收到的banked reset，今天新注册的两个plus账号在4小时后的23:05也都收到了。Tibo没有食言。

[pluoki @pluoki]

@wey_gu 老账号收到了。新注册的两个plus账号都没都到，被晃点了。

[Polymarket @Polymarket]

BREAKING: It’s been revealed NVIDIA's $12,930,300,000.00 acquisition of Hugging Face contains an easter egg.

The first 6 numbers of the acquisition price represent the decimal conversion of Unicode character U+1F917.

The 🤗 emoji.