以下部分引用 AI 总结，现阶段 AI 仍然有幻觉。请以内容的原文为准。

数据新鲜度群聊 07-14 ✓卖方 07-14 ✓Wrap 断档15天 ✕News 07-15 ✓X 断档12天 ✕页面生成 07-15

doodlestein

58 atoms · 跨 22 天 · 首见 2026-04-16 · 最近 2026-07-03

三色: 🟦 fact 25 · 🟥 take 33 · stance ▲16/▼3/◆23

叙事状态: 🍂 fading (退潮) · 💤 沉寂 · 策展近14d 0 atom · 🐦 X 近14d 4

状态只算低频策展源 (群/卖方); X firehose 仅作背景音量

来源: X 58

时态: fresh:50 · aging:7 · stale:1

标签: 好观点:26 · 好数字:19 · 好思考:8 · 好信源:7 · 好问题:1

🟨 AI 综合 · junior analyst 概览

展开 AI 综合 (灰色 · 非市场结论 · 点击数字溯源到原 atom)

model: deepseek-chat · 2026-07-12 · 默认折叠

FrankenSQLite 项目已实现与 C SQLite 的速度和正确性持平，并在几个重要案例上超越后者 ¹ ¹，项目方预期一个月内关闭所有剩余性能差距 ¹。该项目的进展被视为其他 FrankenSuite 项目的标准 ¹，而关于其性能的负面报道被指为不诚实的假新闻 ¹。应用开发者通过 tmux 复用使每个会话与正常交互无区别，并规避账单变更等规则 ¹ ¹。然而，bearish 观点指出多数新创公司资本不足将失败 ¹，且存在使用个人账号规避 API 收费的现象 ¹。账户分布方面，OpenAI/Anthropic 账户各约一半 ¹，持有 54 个账户、22 个 Claude Max、11 个 GPT Pro 和 4 个 Gemini Ultra ¹ ¹ ¹ ¹，月度订阅成本达 12,500 美元 ¹。

🟦 客观事实 (facts) (25)

<a id="atom-f3dc7260da9f1af4"></a>🟦 2026-06-21 X·@doodlestein · account split within openai_anthropic · 账户数
About half and half.
tags: 好数字 · → daily
value: qty=27 / 27 · date=2026-06-21
<a id="atom-310d9a46292c8a02"></a>🟦 2026-06-20 X·@doodlestein · account count openai_anthropic · 账户数
I’m ready to start adding to my 54 accounts
tags: 好数字 · → daily
value: qty=54 · date=2026-06-21
<a id="atom-c8f7fcb7e5cd9944"></a>🟦 2026-06-11 X·@doodlestein · number of Claude Max accounts · 产品订阅
22 Claude Max accounts
tags: 好数字 · → daily
value: qty=22
📷 原图
<a id="atom-a235500570d888fe"></a>🟦 2026-06-11 X·@doodlestein · number of GPT Pro accounts · 产品订阅
11 GPT Pro accounts
tags: 好数字 · → daily
value: qty=11
📷 原图
<a id="atom-7c836b8f53ce83b2"></a>🟦 2026-06-11 X·@doodlestein · number of Gemini Ultra accounts · 产品订阅
my 4 Gemini Ultra accounts
tags: 好数字 · → daily
value: qty=4
📷 原图
<a id="atom-54d0c3aed063a021"></a>🟦 2026-06-09 X·@doodlestein · 生产环境服务方式 · 资本开支
That's just for serving prod, I do the heavy lifting and computationally intensive stuff locally in Postgres. It's not that expensive at my usage (under $100/month)
tags: 好数字 · → daily
value: qty=under $100/month
<a id="atom-de9f671fad945681"></a>🟦 2026-06-01 X·@doodlestein · spends $12.5k/month on LLM subscription costs · 仓位情绪
I can swing the $12.5k/month, but not $200k/month.
tags: 好数字 · → daily
value: qty=$12.5k/month
<a id="atom-e973c506352dc5dc"></a>🟦 2026-06-01 X·@doodlestein · 持有 GPT Pro 订阅数量 · 仓位
I have 22 of them.
tags: 好数字 · → daily
value: qty=22
<a id="atom-059fe264308a3731"></a>🟦 2026-05-31 X·@doodlestein · uses 24 Claude Max accounts and 22 GPT Pro accounts · 仓位情绪/产品使用
I max out the limits of 24 Claude Max accounts and 22 GPT Pro accounts.
tags: 好数字 · → daily
<a id="atom-7c2ec4537c1b643d"></a>🟦 2026-05-27 X·@doodlestein · GPT-5.5和Opus 4.7使用量每周达数十亿tokens · 技术路线
billions of tokens per week with GPT-5.5 and Opus 4.7
tags: 好数字 · → daily
value: qty=billions of tokens per week
📷 原图
<a id="atom-8e6a2d7f994eba98"></a>🟦 2026-05-27 X·@doodlestein · 新发布技能/running-the-gauntlet-on-your-rust-port大小为3.6MB、364个文件 · 产品发布
weighing in at 3.6 megabytes and 364 files (including 63 subagents and dozens of scripts)
tags: 好数字 · → daily
value: qty=3.6 megabytes and 364 files
📷 原图
<a id="atom-c13167f71bf6af11"></a>🟦 2026-05-13 X·@doodlestein · MRR crossed $8k · 营收
I just crossed $8k in MRR in like 9 weeks or so
tags: 好数字 · → daily
value: qty=$8k
<a id="atom-7e45f0d3ba1a516b"></a>🟦 2026-05-13 X·@doodlestein · MRR growth rate · 营收
it grows 1-3% per day
tags: 好数字 · → daily
value: qty=1-3% per day
<a id="atom-480895d8fe04f39c"></a>🟦 2026-05-05 X·@doodlestein · uses Resend API for outgoing emails · 技术细节
Yes, I used Resend API for outgoing emails
tags: 好信源 · → daily
<a id="atom-203643f5e278497f"></a>🟦 2026-05-05 X·@doodlestein · uses Cloudflare for inbound emails · 技术细节
Cloudflare for inbound emails to support.
tags: 好信源 · → daily
<a id="atom-5f69af2d84b6a7d1"></a>🟦 2026-05-02 X·@doodlestein · AI 工具套件定价 · 用户成本
skills are just $20/month
tags: 好数字 · → daily
value: qty=$20/month
<a id="atom-3e0a7cac6d19c3db"></a>🟦 2026-05-02 X·@doodlestein · AI 订阅账户数量 · 用户成本
cranking away with my 52 AI subscriptions
tags: 好数字 · → daily
value: qty=52
<a id="atom-961a4c27cea20efa"></a>🟦 2026-05-02 X·@doodlestein · 社区规模 · 社区生态
nearly 1,000 people in the Agent Flywheel Discord
tags: 好数字 · → daily
value: qty=nearly 1,000
<a id="atom-ab0ea004332cea55"></a>🟦 2026-06-29 X·@doodlestein · already connected with a big owner from this post · 合作客户
You’d be surprised, I’ve already connected with a big owner from this post!
tags: 好信源 · → daily
<a id="atom-496d73ee4be87edc"></a>🟦 2026-06-02 X·@doodlestein · 与Anthropic人员关系 · 人员关注
A lot of Anthropic people follow me, too.
tags: 好信源 · → daily
value: qty=很多关注
<a id="atom-fa65ae4fe352ca79"></a>🟦 2026-06-01 X·@doodlestein · 与OpenAI Codex团队关系 · 团队关注
Half the codex team follows me.
tags: 好信源 · → daily
value: qty=一半成员关注
<a id="atom-ab2d44e9f3e736f9"></a>🟦 2026-05-05 X·@doodlestein [老化中] · 发布Agent工具套件 · 产品发布
I have the most advanced set of open-source, provider-agnostic agent tooling to do all that at scale:
tags: 好信源 · → daily
📷 原图
<a id="atom-027876623773bb24"></a>🟦 2026-06-21 X·@doodlestein · account management method for multiple accounts · 账户管理
Just separate Google accounts for each.
tags: 好思考 · → daily
<a id="atom-d5e2e4ce983c3f31"></a>🟦 2026-06-01 X·@doodlestein · 代码产出 · 产品使用
And I write a lot of code
tags: 好观点 · → daily
value: qty=大量
<a id="atom-28f149269576b6f4"></a>🟦 2026-05-30 X·@doodlestein · 建议使用Hashi Vault管理secrets · 安全措施
Hashi vault for secrets and you’re good.
tags: 好观点 · → daily

🟥 多头 takes (bullish) (12)

<a id="atom-e4765e4f04571aa3"></a>🟥 2026-05-28 X·@doodlestein · 关于FrankenSQLite性能的负面报道是不诚实的假新闻 · 竞争格局
That was always fake news from unscrupulous and dishonest haters
tags: 好观点 · → daily
<a id="atom-571b86fcec386757"></a>🟥 2026-05-28 X·@doodlestein · Rust语言通用性强，技能应能适用于任何Rust代码 · 技术路线
Rust is Rust, it should work well on anything
tags: 好观点 · → daily
<a id="atom-a3131b5f8a55eb93"></a>🟥 2026-05-27 X·@doodlestein · FrankenSQLite项目达到与C SQLite速度和正确性的同等水平 · 产品发布
I recently posted on here about the 3-month campaign I waged in my FrankenSQLite project to bring it up to parity with the C SQLite project in terms of speed and correctness across a wide range of real-world workloads
tags: 好观点 · → daily
📷 原图
<a id="atom-286879fe6404c01a"></a>🟥 2026-05-27 X·@doodlestein · FrankenSQLite在几个重要案例中超越C SQLite · 竞争格局
beyond it in several important cases
tags: 好观点 · → daily
📷 原图
<a id="atom-47d1111ac6c284e8"></a>🟥 2026-05-27 X·@doodlestein · tmux复用方式使每个会话与正常交互会话无区别 · 技术路线
each session is indistinguishable from a normal interactive session
tags: 好观点 · → daily
<a id="atom-3dffc622bf00c2d3"></a>🟥 2026-05-27 X·@doodlestein · tmux复用方式避免违反Anthropic账单变更等规则 · 技术路线
you don’t run afoul of Anthropic billing changes and other nonsense
tags: 好观点 · → daily
<a id="atom-211db7901e5ae4cc"></a>🟥 2026-05-27 X·@doodlestein · FrankenSQLite是其他FrankenSuite项目的标准 · 竞争格局
whatever I do in FrankenSQLite is sort of the standard I’m shooting for across all of them
tags: 好观点 · → daily
<a id="atom-929f1ba9bdd68cd2"></a>🟥 2026-05-08 X·@doodlestein · FrankenSQLite completion timeline · 产品开发
I believe I will be able to close out all remaining performance gaps within the next month or so
tags: 好观点 · → daily
value: date=next month or so
📷 原图
<a id="atom-95e4d032173e651f"></a>🟥 2026-05-08 X·@doodlestein [老化中] · FrankenSQLite completeness · 技术论证
This project is really a kind of existence proof: if I can do this, then I can do something similar for just about ANY piece of software
tags: 好观点 · → daily
📷 原图
<a id="atom-96ca1ce3980aa591"></a>🟥 2026-05-08 X·@doodlestein · FrankenSQLite correctness status · 产品质量
I will also fix every single correctness issue
tags: 好观点 · → daily
📷 原图
<a id="atom-ddf0d406c22f90b0"></a>🟥 2026-05-05 X·@doodlestein · 幸存公司将实现更高单人营收效率 · 营收效率/Agent/替代人力
the ones that survive will be far more efficient in terms of revenue per employee
tags: 好观点 · → daily
<a id="atom-1236d4906b10e1e1"></a>🟥 2026-05-05 X·@doodlestein · 对GDP有好处但对就业无益 · GDP/就业/Agent
good for GDP but not for jobs
tags: 好观点 · → daily

🟥 空头 takes (bearish) (3)

<a id="atom-1dbff6f680e7a934"></a>🟥 2026-06-25 X·@doodlestein · 指出中国已有农场使用个人账号规避限制 · 商业实践/监管套利
There are already farms that do this in China; they have seasoned Google accounts, US cell numbers, virtual credit cards, etc
tags: 好信源·好思考 · → daily
<a id="atom-9bf777afcab32e35"></a>🟥 2026-06-25 X·@doodlestein · 认为对方没有使用API并按API价格付费 · 定价权/商业模式
They absolutely didn’t use the API and pay API rates
tags: 好观点·好思考 · → daily
<a id="atom-f5f587dc48abcfc5"></a>🟥 2026-05-05 X·@goldstein_aa · 多数新创公司资本不足将失败 · 新创公司/失败风险
most of those new companies won’t be adequately capitalized and will fail
tags: 好观点 · → daily

🟥 中性 takes (neutral) (15)

<a id="atom-6389b4cbd79e0d1b"></a>🟥 2026-06-22 X·@doodlestein · browser profile management · 成本评估/账户管理
It’s really not, you log in at the browser level. I use chrome, chrome canary, and brave. You know what’s a nightmare? Switching to the API and paying $200k/month instead for my usage. This is a minor inconvenience in comparison.
tags: 好数字·好观点 · → daily
<a id="atom-8995877700cf57a2"></a>🟥 2026-06-22 X·@doodlestein · api cost comparison vs consumer subscription · 成本评估
Switching to the API and paying $200k/month instead for my usage.
tags: 好数字 · → daily
value: qty=$200k/month
<a id="atom-3749c423747fe960"></a>🟥 2026-06-21 X·@doodlestein · infrastructure cost estimate for high volume local deployment · 成本评估
Correct. I’d need a whole datacenter costing many millions a year.
tags: 好数字 · → daily
value: qty=many millions per year
<a id="atom-33f8ec74365deaea"></a>🟥 2026-06-23 X·@doodlestein · 对财富层级看法的条件判断 · 仓位情绪
@trentjhughes Totally depends on geography and number of kids.
tags: 好思考 · → daily
<a id="atom-8e4f3536045a8801"></a>🟥 2026-05-05 X·@doodlestein · 许多公司将成为少数人加海量Agent团队 · 劳动结构/Agent/替代
many will likely be a few dudes and hordes of agents
tags: 好思考 · → daily
<a id="atom-018e6b5d7e85ea45"></a>🟥 2026-04-18 X·@doodlestein [老化中] · 对davidshor的转发回复表示同意 · 仓位情绪
Yup: https://t.co/gW2BVZ6NDZ
tags: 好思考 · → daily
<a id="atom-8af22672fe406bd6"></a>🟥 2026-04-16 X·@doodlestein [老化中] · approach to clean room re-implementation project · 技术路线
approach it almost like a clean room re-implementation project
tags: 好思考 · → daily
<a id="atom-7003070036d63deb"></a>🟥 2026-04-16 X·@doodlestein [老化中] · thinking about best approach for new skill · 产品发布
I will think about the best way to approach it
tags: 好思考 · → daily
<a id="atom-7111745c98144af9"></a>🟥 2026-07-03 X·@doodlestein · 观点--进入对冲基金通常问题在于技能/知识而非人脉或学历 · 仓位情绪
It's usually a skill/knowledge problem
tags: 好观点 · → daily
<a id="atom-9ae2960e04662aa0"></a>🟥 2026-07-03 X·@doodlestein · 声称自己可通过技能和几小时工作超越测试结果 · 个人能力
I am positive I could outperform these guys with a skill and a few hours of work.
tags: 好观点 · → daily
<a id="atom-88b64a99574ccf59"></a>🟥 2026-06-11 X·@doodlestein [老旧] · 预测政治逻辑不可抗拒 · 地缘政治
Yup, this coming, whether it’s a good idea or not. The political logic is too irresistible. I’ve been predicting this for a while
tags: 好观点 · → daily
<a id="atom-fc3eeadd74655d48"></a>🟥 2026-05-31 X·@doodlestein · 使用多模型策略观点 · 仓位情绪
I just always used both and think it’s the only sensible approach if you want the best results and aren’t budget constrained
tags: 好观点 · → daily
<a id="atom-52f8d1eac4648910"></a>🟥 2026-05-28 X·@doodlestein · 不愿花时间在其他人的项目上 · 合作客户
I’ve learned my lesson about spending my time and tokens on other people’s projects!
tags: 好观点 · → daily
<a id="atom-f3c59f952aa1a8ac"></a>🟥 2026-05-27 X·@doodlestein · 所有工作设计为使用Claude Max和GPT Pro订阅本地运行 · 技术路线
Everything I do is designed to work locally using Claude Max and GPT Pro subscriptions
tags: 好观点 · → daily
<a id="atom-98d4b59155f977f5"></a>🟥 2026-04-16 X·@doodlestein [老化中] · making a skills exchange feature · 产品发布
I'm making this now
tags: 好观点 · → daily

⏱ 时间轴 (近 20)

🟥 2026-07-03 · narrative · 观点--进入对冲基金通常问题在于技能/知识而非人脉或学历 · → daily
🟥 2026-07-03 · narrative · 声称自己可通过技能和几小时工作超越测试结果 · → daily
🟥 2026-07-03 · narrative · doing · → daily
🟦 2026-06-29 · fact · already connected with a big owner from this post · → daily
🟥 2026-06-25 · narrative · 认为对方没有使用API并按API价格付费 · → daily
🟥 2026-06-25 · fact · 指出中国已有农场使用个人账号规避限制 · → daily
🟥 2026-06-23 · refute · 对财富层级看法的条件判断 · → daily
🟥 2026-06-22 · narrative · browser profile management · → daily
🟥 2026-06-22 · forecast · api cost comparison vs consumer subscription · → daily
🟥 2026-06-21 · forecast · infrastructure cost estimate for high volume local deploymen · → daily
🟦 2026-06-21 · fact · account split within openai_anthropic · → daily
🟦 2026-06-21 · fact · account management method for multiple accounts · → daily
🟦 2026-06-20 · fact · account count openai_anthropic · → daily
🟦 2026-06-11 · fact · number of Claude Max accounts · → daily
🟦 2026-06-11 · fact · number of GPT Pro accounts · → daily
🟦 2026-06-11 · fact · number of Gemini Ultra accounts · → daily
🟥 2026-06-11 · position · 预测政治逻辑不可抗拒 · → daily
🟦 2026-06-09 · fact · 生产环境服务方式 · → daily
🟦 2026-06-02 · fact · 与Anthropic人员关系 · → daily
🟦 2026-06-01 · fact · spends $12.5k/month on LLM subscription costs · → daily

← 实体目录 · 系统日志