Sabitlenmiş Tweet
Sibjax
527 posts
Sibjax
@sibjax
Building agentic GTM
San Francisco, California Katılım Nisan 2025
182 Takip Edilen98 Takipçiler
@seanboisselle Evan is a moral guy - he’s chosen his employees and family over bottom line many many times over the years
I wouldn’t be so quick to dismiss him
English
Working for FAANG isn't what it used to be.
The bar for hiring is so high, but you're still a disposable cog.
Freelance / entrepreneurship is the way.
Bloomberg@business
Snap is laying off roughly 1,000 full-time employees, or 16% of its global workforce, part of an effort by CEO Evan Spiegel to reduce costs and achieve profitability bloomberg.com/news/articles/…
English
@AYi_AInotes @Teknium But obviously the timeline is wrong since Hermes was first pushed 2025….
So maybe you’re the one that’s wrong, right?
English
本来今天要写一篇Hermes-agent相比小龙虾token消耗更大的推文，
但刷到了这个在程序员圈和开源社区炸了的瓜，
我花了两个多小时扒完两边的仓库和证据链，说实话越看越后背发凉，
不是因为抄袭本身，而是这可能是AI时代第一起被完整记录的架构级代码洗稿：
没有复制一行代码，文本相似度0%，但核心架构的同构度几乎是100%，
我尽量从技术角度把前因后果拆清楚，具体兄弟们你们自己判断。
先说时间线，这是整件事的基础，所有时间戳都来自GitHub仓库元数据，任何人都可以去验证，
2月1日，EvoMap团队开源了Evolver，一个AI Agent自进化引擎，核心是他们自研的GEP协议，10分钟登顶ClawHub热榜，
到2月16日，整套协议体系通过多篇公开文章全部公开：包括Gene/Capsule/Event三级资产体系、Scan-Select-Mutate-Validate-Solidify进化循环、信号选择器、反射机制、叙事记忆，全部摊在了桌面上，
3月9日，Nous Research创建了hermes-agent-self-evolution仓库，
3月12日发布v0.2.0正式推出完整的技能生态系统，
中间隔了24到39天
时间线只是起点，真正让我震惊的是架构层面的模块级一对一对应，我拎几个最硬的出来，
第一，进化闭环完全同构，Evolver的核心循环是任务完成后自动提取可复用资产并持久化，Hermes的官方描述是Task completes → Agent evaluates → writes SKILL.md → Future tasks load automatically，范式一模一样，只是Evolver用Gene/Capsule的JSON结构，Hermes用SKILL.md的Markdown结构，
第二，三层记忆体系精确对齐，
Evolver有EVOLUTION_PRINCIPLES.md（持久事实）+ Gene/Capsule JSON（程序性记忆）+ events.jsonl（历史搜索），Hermes有MEMORY.md + USER.md（持久事实）+ SKILL.md文件（程序性记忆）+ SQLite FTS5（历史搜索），不是两层不是四层，精确的三层，且每层的语义角色一一对应，
第三，周期性反射机制，
Evolver每5个进化周期触发一次战略性自我评估，Hermes每15次tool call运行一次self-evaluation checkpoint，目的完全一致：从执行经验中提取模式并持久化。
这还没完，两个项目的进化主循环都是10步编排，
Evolver是ensureAssetFiles → extractSignals → getMemoryAdvice → selectGene → buildMutation → selectPersonality → buildPrompt → writeArtifact → writeState → reflect，
Hermes是find_skill → build eval set → baseline validate → config optimizer → GEPA optimize → extract text → evolved validate → holdout eval → report → save，
核心模式完全一致——加载 → 评估 → 选择/优化 → 验证 → 持久化，
更关键的是源码模块的一对一对应，
Evolver的selector.js对应Hermes的skill_commands.py，solidify.js对应skill_manager_tool.py，reflection.js对应每15次tool call自评估，memoryGraph.js对应memory_tool.py，skillDistiller.js对应evolve_skill.py，executionTrace.js对应trajectory.py，
我数了一下，Evolver的11个核心模块，Hermes每一个都有功能等价的对应文件
有人可能会问，会不会只是英雄所见略同，两个团队独立做出了相似的设计？
说实话如果只是单一维度的相似，我不会花几个小时研究和写这条推文，从经验中学习本身就是通用AI概念，周期性自评估在学术界也有先例，
但问题在于：三层记忆体系、三级资产结构、10步进化循环、运行时渐进式技能发现、多维加权适应度评分、原子写入、安全扫描、注入防护、容量控制，这些选择在同一个项目中、同一个时间窗口内同时收敛的概率，随着每多一个维度的匹配呈指数级递减，
而且最关键的一点是对Hermes两个仓库做全文搜索，EvoMap、evolver、Genome Evolution Protocol、capsule、solidify、signals_match，全部零匹配，没有任何代码残留，这恰恰符合AI跨语言重写的特征：AI重写架构时不会保留原项目的特征性字符串，但架构层面的同构性无法被重写消除。
然后说说双方的回应，
Hermes Agent昨天下场回复了，大意是说他们的仓库2025年7月22日就创建了，比Evolver还早，但这里有个关键事实：
那个仓库在2026年2月25日之前一直是私有项目，v0.1.0自己都写着叫initial pre-public foundation，技能生态系统直到3月12日的v0.2.0才正式发布，没有任何公开证据能证明他们在私有阶段已经包含了自进化能力，
更耐人寻味的是，这条回复后来被秒删了，Evolver创始人也被拉黑了，
另外要说一个公平的点，Hermes的自进化仓库用了GEPA这个来自Berkeley/Stanford的独立学术成果，是正当的技术选型，Anthropic的Agent Skills标准也早于Evolver，Hermes采用SKILL.md格式是合理的行业选择，但这些都不能解释整体架构层面的高度同构，
开源社区有个基本惯例，LangChain引用了DSPy，CrewAI对比了AutoGen，MetaGPT引用了相关多agent框架，发现同领域先行项目时加一句Related Work是标准做法，而Hermes在7份公开材料中对Evolver只字未提。
说实话这件事让我想了很久的一个问题是：
AI时代的代码洗稿要怎么防？
传统的查重工具看的是文本相似度，但现在AI可以把你的整套架构吃透，换一个语言从Node.js变Python，换一套术语Gene变SKILL.md、solidify变skill_manage，调整一下文件结构，吐出一个文本相似度0%但架构DNA完全一致的产物，
这不是个案，今年已经接连发生了好几起：
美团Tabbit AI源码残留原项目名称，
三省六部AI朝廷开源21小时后被AI重写文本相似度仅3%但15个核心设计全部一致，
微软Peerd复制个人开源项目Spegel代码，
EvoMap团队最后的选择是把协议从MIT改成GPL，核心模块改为混淆发布，说实话我能理解但也觉得很心酸，
他们原话是：别人用AI洗得走代码，但洗不走我们对下一步路径的认知，洗不走这几个月踩坑换来的直觉，
这话没毛病，但如果开源意味着你的心血在几周内就被资源更多的团队用AI洗成他们的首创，谁还愿意做那个开荒的人？
这个问题没有答案，但值得每个开发者认真想想。
autogame-17@autogame_17
We @EvoMapAI spent months and countless sleepless nights building Evolver. A well-resourced team behind Hermes Agent "reinvented" it in just 30 days. ● Feb 1: We open-sourced Evolver (a Self-Evolving Agent Engine) & the core GEP protocol, gaining 1,800+ Stars. ● Mar 9: Hermes Agent hastily created their repo and launched. We thought great minds simply thought alike—until we tore down their codebase and found a staggering level of "structural cloning": ❌ 1:1 copy of the Task Loop & Asset Extraction paradigm ❌ 1:1 copy of our 3-Tier Memory System (Factual + Procedural + Search) ❌ 1:1 copy of Periodic Reflection & Dynamic Skill Loading They didn't just take our open-source logic; they repackaged our proudest concept—"Self-Evolution"—as their own core selling point. Took everything. Zero attribution. Big teams might have louder megaphones, but commit timestamps don't lie. We aren't here to play judge. We're just putting the code comparisons on the table. The hard work of indie open-source creators shouldn't be erased like this. Full architectural breakdown and code evidence 👇: evomap.ai/blog/hermes-ag…
中文
"sf but with hot people" is an oxymoron if i've ever seen one
Christy Lam@christyoverflow
is miami the next sf but with hot people
English
🚨 TRUMP PRESS SEC. KAROLINE LEAVITT PRAISES PAKISTAN!
"I just want to make one point that's important to the President. The Pakistanis have been incredible mediators throughout this process and we really appreciate their friendship and their efforts to bring this deal to a close!"
"So they are the only mediator in this negotiation while there have been many countries around the world who want to offer their help."
"The President feels it's important to continue to streamline this communication through the Pakistanis and so that's what continues to take place."
English
@icanvardar If you were Anthropic and you had AGI wouldn’t you also need the heck out of the publicly available models and use it for yourself?
English
codex has been more stable lately. claude somehow getting worse every week
English
I’ve thought about this a lot, Discord for business would double their revenue. Know so many founders who would use it if discord was SOC2 compliant
saila@sailaunderscore
This is an open letter to @jasoncitron and the board at @discord. cc: @benchmark @IndexVentures @Greenoaks @GreylockVC @coatuemgmt @Sony @TencentGlobal Discord is an amazing product. It offers freemium private or public servers for communication, these can be used as forums, information caches, or more. Discord’s problem is well-known, it sells to a younger userbase, this creates a wide series of problems — these users are fickle, they expect more, and they do not expect to pay — contrast this to other app-based private forums like Slack where users: pay (a lot! see attached), ask for relatively little, and do not churn off. Slack was purchased by Salesforce in 2021 for $28bb, it’s estimated current revenue is around $2.5bb. Discord is a private company with an estimated current valuation of around $7-10bb, and revenue of around $800mm. Interestingly, a Discord user would never consider using Slack. A Slack user would consider using Discord. The largest problem impeding this is a branding problem, Discord is used primarily by online-first and young communities, such as gamers. Not particularly attractive branding for an enterprise company looking to have a server for communications. Meanwhile both Slack itself and Discord users would never consider the other, because of a financial mismatch. With that said, the Discord product is fantastic and contains everything needed on a basic level to compete with Slack. The solution is exceedingly simple, Discord needs to hire a semi-autonomous team, which has access to their core IP for a product whitelabel. If Discord was willing to empower a small team to create, say “Accord, powered by Discord”, it could immediately aim to double its market capitalization and attack Slack’s userbase. Humam Sakhnini, your current CEO is a former consultant and investment banker, he is optimized for CFOing a company to public markets to engineer an exit. With the rise of vibe-coding and LLMs, SaaS multiples have cratered and the private secondary valuation of Discord has fallen steeply. Discord needs a CEO who can empower teams for asymmetric upsides, such as attacking a new, better-paying demographic, with a very small team dedicated entirely to whitelabelling and rebranding Discord’s product for enterprise. I hope you can rally the team and think beyond simply iterating endlessly on minor features and consider louder and more interesting iterations to Discord as an entity. I am happy to stand-in as an external consultant in a pro-bono capacity. Best regards, Saila
English
@weswinder @sergeynazarovx If anything we’re getting more authentic
Counter intuitive
English
i don’t even know what they do
but wouldn’t be surprised if the answer was fraud somehow
people are too focused on flexing arr for some reason
Akimasa Hayashi@aki_hayashii
English
@ShengXiaoy15782 @Steve_Yegge It was a small sample like I said, maybe your org is different but there was definitely a sense of superiority expressed by "senior" engineers
English
@sibjax @Steve_Yegge I don't know how true this statement is. Everyone in my team is using AI to generate code. At senior engineer level, nobody evaluates your performance by the code you write. Senior engineers are evaluated by the areas they drive and the impacts they make.
English
I was chatting with my buddy at Google, who's been a tech director there for about 20 years, about their AI adoption. Craziest convo I've had all year.
The TL;DR is that Google engineering appears to have the same AI adoption footprint as John Deere, the tractor company. Most of the industry has the same internal adoption curve: 20% agentic power users, 20% outright refusers, 60% still using Cursor or equivalent chat tool. It turns out Google has this curve too.
But why is Google so... average? How is it that a handful of companies are taking off like a spaceship, and the rest, including Google, are mired in inaction?
My buddy's observation was key here: There has been an industry-wide hiring freeze for 18+ months, during which time nobody has been moving jobs. So there are no clued-in people coming in from the outside to tell Google how far behind they are, how utterly mediocre they have become as an eng org.
He says the problem is that they can't use Claude Code because it's the enemy, and Gemini has never been good enough to capture people's workflows like Claude has, so basically agentic coding just never really took off inside Google. They're all just plodding along, completely oblivious to what's happening out there right now.
Not only is Google not able to do anything about it, they don't seem to be aware of the problem at all. I'm having major flashbacks to fifty years ago as a kid at the La Brea Tar Pits, asking, "why can't they just climb out?"
My Google friend and I had this conversation over a month ago. I didn't share it because I wanted to look around a bit, and see if it's really as bad as all that. I've been talking to people from dozens of companies since then. And yeah. It's as bad as all that.
Google is about average. Some companies at the bottom have near-zero AI adoption and can't even get budget for AI. They may have moats and high walls, but the horde is coming for them all the same.
And then there are a few companies I've met recently who are *amazingly* leaned in to AI adoption. One category-leader company just cancelled IntelliJ for a thousand engineers. That's an incredibly bold move, one of many they're making towards agentic adoption. In my opinion, that company is setting themselves up for a _huge_ W.
As for the rest, well, it's the Great Siloing. Everyone's flying blind. With nobody moving companies, no company knows where they stand on the AI adoption curve. Nobody knows how they're doing compared to everyone else.
Half of them just check a box: "We enabled {Copilot/Cursor} for everyone!" Cue smug celebrations. They think this is like getting SOC2 compliance, just a thing they turn on and now it's "solved." And they don't realize that they've done effectively nothing at all.
All because of a hiring freeze.
English
@_NathanCalvin "AI did it!" is quickly becoming the new "Dog ate my homework"
English
The CEO of Krafton (creator of PUBG) asked ChatGPT to create a "corporate takeover strategy" to prevent a company they acquired from hitting a revenue target within a certain time window (which would trigger an additional payout).
ChatGPT (against his lawyer's advice) suggested locking down the acquired companies Steam account to prevent them from publishing Subnautica 2 in the time window, which the CEO of Krafton followed.
ChatGPT's advice did not hold up at trial and the judge was not happy. The opinion is a wild read and includes several direct quotes from the Krafton CEO's ChatGPT conversation.
I feel like it's gonna take a few more high profile examples like this until executives start realizing that conversations with ChatGPT are not privileged and you probably shouldn't describe your questionably legal schemes to them in detail!
English
And the other reality is that my friends at Square worked 2 hours a week from whatever remote location they wanted too
You hear stories about engineers that have really low hours and Square / Block surpassed all of them in terms of being able to chill
unusual_whales@unusual_whales
Jack Dorsey of Block has said that AI can make middle management obsolete, per FORTUNE.
English
@valdombre No country is ever single handedly ruined
Anyone telling you that is wrong
English
This guy single-handedly ruined a G7 nation's economy and social fabric and then retired and took his girlfriend Katy Perry to Coachella
English