Filter
Exclude
Time range
-
Near
Meta Platforms 计划通过限制员工的 Token 使用额度,来管控公司内部飙升的 AI 成本。根据一份泄露的内部备忘录,Meta 正在构建一个名为 AI Gateway 的中央网关,用于实时监控员工的 AI 用量与支出、设定预算并对 Token 消耗设置上限。Meta 预计在 2026 年,仅内部 AI 使用就将产生数十亿美元的费用。 限额举措与 Meta 此前极力推广 AI 的态度形成鲜明反差。2025 年 11 月,Meta 曾通知员工将展示「AI 驱动的影响力」作为 2026 年的核心考核标准,并将绩效奖金与 AI 使用率挂钩。过度推广引发了员工间竞相刷量的「tokenmaxxing」狂欢,甚至一度出现名为「Claudeonomics」的内部排行榜来公开展示用量排名。在排行榜被关停前,员工在 30 天内消耗的 Token 总量曾飙升至 73.7 万亿。Meta 首席技术官 Andrew Bosworth 随后发出警示,强调单纯增加 Token 消耗并不代表产出,员工应当在能切实提升效率的场景下使用 AI 工具。 为了进一步缩减支出,Meta 开始将内部 AI 开发重心向自研工具倾斜。泄露的备忘录显示,Meta 正在推动员工逐步停用 Anthropic 旗下 Claude 等第三方编程工具,转而使用自研的编程助手 MetaCode(曾用名 Devmate)。Meta 新成立的应用 AI 工程部门已被要求全力改进 MetaCode,通过生成编程挑战来生产高强度的强化学习训练数据。尽管 Meta 仍允许员工访问外部模型,但未来将在自建网关中实施更严格的预算与配额审批机制。 Meta 并非唯一因大模型用量过载而面临财务压力的企业。在 2026 年初,Uber 和 ServiceNow 等公司在短短几个月内便耗尽了全年的 Anthropic 额度。ServiceNow 已对员工实施日用量监测,部分风险投资机构也开始对内部 AI Token 使用设定日均消费上限,以防止无节制的算力成本扩张。
After encouraging staff to prove their “AI-driven impact,” Meta is now moving to cap employee token usage and steer workers toward in-house tools. Full story: thein.fo/4vbt2PE
1
1
1
915
메타, AI 사용 비용 급증에 따라 토큰 사용량 제한 방안 추진 : 내부 메모에서 직원 AI 사용량 및 비용을 실시간 추적하고 예산 설정·토큰 사용량 한도 관리를 위한 내부 플랫폼 구축 시사 "AI 사용량이 기하급수적으로 증가하고 있으며, 2026년에는 내부 사용만으로도 수십억 달러를 지출할 것으로 예상된다. 동시에 개인과 팀은 AI 사용 방식에 대한 가시성과 통제력이 제한적이다. 2027년부터는 예산, 자원 배분 결정, 지원 도구 등을 기반으로 보다 체계적으로 AI 토큰을 관리하게 될 것으로 예상한다." : 비용 관리 강화를 위해 개발자 및 엔지니어 조직은 AI Gateway라는 중앙 대시보드 구축. 사용량·비용 모니터링, 예산 관리, 이상 사용 탐지 등의 기능 제공 예정 : 코딩 업무에서 외부 AI 도구 사용을 축소하고 자체 솔루션 확대 추진. 자체 개발 코딩 어시스턴트 MetaCode(구 Devmate)가 대표적 : 신규 조직 Applied AI Engineering(AAI) 팀이 MetaCode 성능 개선 작업 진행 : 엔지니어들에게 프로그래밍 문제를 생성하고 MetaCode가 이를 해결하도록 하는 방식의 강화학습(RL) 데이터 구축 지시 : AI 사용 장려 정책 이후 Tokenmaxxing 경쟁 발생. Claudeonomics라는 내부 리더보드에서 경쟁했으며, 4월 기준 30일간 총 60.2조 개 토큰 사용. 이후 73.7조 토큰까지 증가하며 해당 리더보드는 폐쇄 : 메타 CTO Andrew Bosworth 는 토큰 사용량 자체는 성과 지표가 아니라며 과도한 AI 사용 자제를 지시 : 동시에 Agent Transformation Accelerator 프로젝트를 통해 사내 업무를 AI 에이전트 중심으로 재편 중 t.me/Samsung_Global_AI_SW
4
389
메타, AI 사용 비용 급증에 따라 토큰 사용량 제한 방안 추진 : 내부 메모에서 직원 AI 사용량 및 비용을 실시간 추적하고 예산 설정·토큰 사용량 한도 관리를 위한 내부 플랫폼 구축 시사 "AI 사용량이 기하급수적으로 증가하고 있으며, 2026년에는 내부 사용만으로도 수십억 달러를 지출할 것으로 예상된다. 동시에 개인과 팀은 AI 사용 방식에 대한 가시성과 통제력이 제한적이다. 2027년부터는 예산, 자원 배분 결정, 지원 도구 등을 기반으로 보다 체계적으로 AI 토큰을 관리하게 될 것으로 예상한다." : 비용 관리 강화를 위해 개발자 및 엔지니어 조직은 AI Gateway라는 중앙 대시보드 구축. 사용량·비용 모니터링, 예산 관리, 이상 사용 탐지 등의 기능 제공 예정 : 코딩 업무에서 외부 AI 도구 사용을 축소하고 자체 솔루션 확대 추진. 자체 개발 코딩 어시스턴트 MetaCode(구 Devmate)가 대표적 : 신규 조직 Applied AI Engineering(AAI) 팀이 MetaCode 성능 개선 작업 진행 : 엔지니어들에게 프로그래밍 문제를 생성하고 MetaCode가 이를 해결하도록 하는 방식의 강화학습(RL) 데이터 구축 지시 : AI 사용 장려 정책 이후 Tokenmaxxing 경쟁 발생. Claudeonomics라는 내부 리더보드에서 경쟁했으며, 4월 기준 30일간 총 60.2조 개 토큰 사용. 이후 73.7조 토큰까지 증가하며 해당 리더보드는 폐쇄 : 메타 CTO Andrew Bosworth 는 토큰 사용량 자체는 성과 지표가 아니라며 과도한 AI 사용 자제를 지시 : 동시에 Agent Transformation Accelerator 프로젝트를 통해 사내 업무를 AI 에이전트 중심으로 재편 중 t.me/Samsung_Global_AI_SW
2
378
$Meta içinde Claudeonomics diye bir liderlik tablosu varmış. Yapay zekayı en çok kullanan çalışanları sıralıyormuş. 85 binden fazla çalışan takip edilmiş, 30 günde 60 trilyon token tüketilmiş. Mühendis başına yıllık AI harcaması 100-300 bin $ arasındaymış. Yani neredeyse bir mühendisin yıllık maaşına denk. Tablonun altındaki çalışanlar ise işten çıkarılma riski taşıyormuş :) (Az kullanan çalışanları çıkarmak için zemin hazırlanıyor gibi geldi bana) $Meta önce rakiplerinin araçlarını kullanıp bilgi sızdırmamak için kendi aracı Devmate i geliştirmiş, çalışanlar beğenmemiş yetersiz bulmuşlar. En son kuralı gevşetmişler, müşteri verisine dokunmadığın sürece rakiplerin aracı Claude Code’u kullanabilirsin demişler. Yapay zeka sayesinde verimlilik şirket için 100 kat artmış ama gelire etkisi 1 kat bile değil. Üstüne 100 milyar $ bandında capex yapan sektörün en iyi AI uzmanlarına milyarlarca dolar bonus dağıtıp ekiplerine katmaya çalışan bir şirket $META, ilginç bir durum.
9
3
105
14,626
Replying to @IndianTechGuide
Internal reports suggest Meta's new tools, like Metamate and DevMate, are now handling up to 70% of routine coding and admin work, reducing the need for manual oversight. Meta isn't alone; so far in 2026, tech layoffs have already surpassed 75,000 globally, led by firms like Amazon and Oracle.
2
280
Meta just showed every CEO in America exactly how to replace their workforce. They laid off 700 people this week. Plan 15,000 more ASAP. Now, leaked internal documents reveal what comes next. Meta now requires 65% of its engineers to write 75% or more of their code using AI by mid-2026. Their Scalable Machine Learning org has a target of 50-80% AI-assisted code. Across Messenger, WhatsApp, and Facebook, 55% of all code changes must be “Agent-Assisted.” 80% of mid-to-senior engineers must adopt AI tools like DevMate and Google Gemini. Starting this year, every Meta employee is graded on “AI-driven impact” in their performance reviews. It is now a core expectation. If you don’t use AI, you don’t advance. Meta is the first major tech company to formally tie promotions to AI adoption. They built an internal game called “Level Up” that rewards employees with badges for hitting AI milestones. They rolled out an “AI Performance Assistant” to help write reviews. They are rebranding engineers as “AI Builders” and reorganizing teams into small “AI-native pods.” Zuckerberg said projects that once required large teams can now be handled by one “very talented” person. Read that again. One person replacing a team. Here’s why this matters beyond Meta. Every CEO in tech watches what Meta does. When Meta cut 11,000 in 2022, the rest of tech followed within months. When Meta tied performance reviews to AI, KPMG did the same thing within weeks. Accenture just told senior staff that AI tool usage will determine who gets promoted to leadership. The playbook is spreading. The pattern is clear. Step one: mandate AI adoption internally. Step two: measure how much output AI handles. Step three: when AI handles 75% of the work, you need 75% fewer people to do it. Step four: cut. Block already did this. Jack Dorsey cut 40% of the company and told shareholders “intelligence tools have changed what it means to build and run a company.” The stock surged 24%. Wall Street doesn’t reward hiring. It rewards headcount reduction. Every earnings call where a CEO says “AI is making us more productive” is a preview of the next layoff announcement. Meta is spending $135 billion on AI this year. They are simultaneously cutting thousands of workers. Those two facts are not in tension. They are the same strategy.
Are you paying attention to what Meta quietly did this week? 700 laid off. 15,000 more expected, as per Reuters. Same week, 6 executives got $921M in stock options at current market cap, and up to 2.7B as it goes higher. 700 lost their jobs. Six got 100s of millions in bonus.
85
331
1,528
954,048
When @jevering first described his vision for @GuildAI and how engineering teams will work in the near future, everything seemed obvious in hindsight. Maturing models and agents mean engineering teams will adopt and deploy everything from 3rd-party agents to homegrown systems, each for a specific task. An agent for deployment. Another for code review. Another for product security. This creates the classic enterprise challenge of managing and maintaining all of them. Guild.ai is building the control plane for AI reliability, observability, and governance at scale. And there's no better team to do it. They built Devmate @Meta, saw firsthand the viral adoption of this paradigm across a scaled engineering org, and most importantly, lived the challenges after that only show up with scale. And this team learns from their customers fast. I've known @jevering for 6 months and I haven't seen another team in the area that ships and reacts to feedback anywhere near as quickly. I'm honored to lead @khoslaventures investment in Guild.ai and to be partnering with James and the team.
Guild.ai has raised a Series A to build the neutral control plane for AI agents. Universal by design. Guild is neutral toward models, vendors, and frameworks. Guild turns agents into shared production infrastructure, with a managed software center for trusted agent capabilities, and an agent hub for discovering and sharing agents. Read the full announcement: guild.ai/knowledge/guild.ai-… Our new site is live: guild.ai
1
1
23
8,703
最近cluade用的我有点恍惚AI ptsd 正好这周也是我们Ai week。我觉得这波 claude对我的改变远大于gpt。尤其对一些本身就卷、工作生活边界本模糊的人来说,AI 时代是更累,更停不下来。上个周末,我几乎全程钉在电脑前。 先分享3个小点吧; 1 曾经看一份技术文档,我会一行一行跟着敲、一步步试。最近就是:先扫一眼确认是自己要的,直接复制链接丢给 AI,输一句”跟着文档走一遍”剩下的自己疯狂按2就完事了。 2.曾经我会幻想 10 年、20 年后的工作和未来。现在脑子里想的,只有半年、一年后我还能吃这碗饭不? 这一波AI 迭代太快了,还记得去年四月份入职的时候devmate还是个智障… 3. 昨天看skills时灵光一闪:我们组 tech lead 审核 diff 极其严,每次给我的 comments 一堆。于是我直接Claude Code 让他帮我整理过去一年他给我留的comments,让它把相似内容整合、按出现频次排序,整理成文档。再让 AI 基于这些,生成自己skills… 现在那些 10x,工程师的通用技巧再好,但很多具体语言框架,组内的一些写法很多时候并不适用。 顺着这个思路:是不是可以把全组同事、甚至隔壁组技术大佬的评论、代码风格、优质写法全部喂给 AI生成skills?是不是可以按语言,框架生成?还能不能再 scale up?还能做什么更极致?最近很难把大脑从ai抽离出来。以至于我自己都分不清:是恐惧在 AI 时代被淘汰,还是真的对这个新世界好奇到停不下来。 感觉我们现在就像闯进一个新的星球。有人喊挖到了黄金,有人叫找到了水源,有人看见了新奇的生物,所有人都在疯狂挥舞锄头尖叫着。但没有人知道,在这个星球里,到底什么才是最珍贵的,我们到底要建起一个怎样的家园。我们还惯性地把地球上珍贵的东西,搬过来当作珍宝,可答案究竟是什么,谁也不知道。 还有一种很真实的感受: 几天不碰 AI、不上班,就觉得自己掉队了,听不懂别人在聊什么。可就算一个月没学,回头看,好像也只落后了一周,因为新的东西,早就把前三周的内容全部取代了。 这种既兴奋又恐慌、既充满希望又极度疲惫的割裂敢最近一直在缠绕着我。哎 艹 #AI#Claude #职场人现状 #技术人日常 #AI焦虑 #工作效率 #内卷#meta码农#码农生活
2
80
Jan 23
Replying to @aadishvdotdev
Devmate, Meta’s internal vscode extension.
2
153
🚨 Anthropic just dropped Claude Opus 4.5, and it's crushing every human on their 2-hour engineering test. Not most humans. EVERY human. The new model is dominating in: Advanced coding & software development Excel & PowerPoint generation Financial modeling & analysis Autonomous AI agents that learn & improve themselves Even Meta is using Claude to power their internal Devmate coding assistant despite being direct competitors in the AI race. This is Anthropic's 3rd major model release in just 2 months. The AI race isn't slowing down; it's accelerating. anthropic.com/news/claude-op…
3
2
8
331
13 Oct 2025
AI Coding Assistants Revolutionize Software Development The latest trend in coding is the rise of AI coding assistants, with Meta's DevMate being a recent example. These tools are transforming software development by scanning codebases, suggesting fixes, and refactoring legacy code, effectively moving AI from assistant to co-creator in the development process. What are your thoughts on the future of AI in coding?
7
6
4,125
6. Meta (Anthropic/Claude Code) There are reports that $META has rolled out an internal coding assistant called "Devmate." which is powered by Anthropic's models. They suggest Meta is embedding Claude internally for dev workflows - Not as a consumer product.
1
3
156
26 Sep 2025
Meta's use of external models: • Vibes: FLUX.1 Kontext Midjourney licensed and live • Ads: In talks with Gemini for targeting and creative use • Meta AI: Exploring OpenAI as a bridge while next Llama trains • Engineering: Claude is used in internal 'Devmate' product for coding If the next Llama misses, huge CapEx spend will face major ROI questions.
1
1
11
3,125
Replying to @TheStalwart
Can't blame this on not having the latest and greatest..Meta's internal AI coding assistant, Devmate, utilizes Claude models, perhaps they are seeing different results?
5
1,361
21 Jun 2025
@MichaelEzeana and I founded Devmate because, like many devs, we wanted to spend more time coding — not writing tests, docs, or READMEs. AI won’t take your job, it just makes the boring parts faster. Join the waitlist devmate.surge.sh/#waitlist

21 Jun 2025
You know how it’s stressfull writing tests, API documentation, or even a README— well, not anymore. Introducing DEVMATE: your AI-powered tool that understands your codebase and automatically generates clear, context-aware API documentation, READMEs, and tests.
4
8
199
21 Jun 2025
You know how it’s stressfull writing tests, API documentation, or even a README— well, not anymore. Introducing DEVMATE: your AI-powered tool that understands your codebase and automatically generates clear, context-aware API documentation, READMEs, and tests.
To wrap up Codelab, we have @0xAgbero for his product demo. An API documentation tool. 🚀
4
10
1,141
Today I learned Meta has an LLM-agent bot (called Devmate) that looks at test failures, diagnoses them, and puts up diffs to fix them. People have been talking about AI doing these types of workflows and whether it's practical or not for quite some time... 1/N
1
1
4
502
after several iterations, I made some changes to devspace (prev. DevMate). need your feedback!!! thoughts?? cc @_buildspace @_nightsweekends #buildinginpublic #s5
I love meeting new people and collaborating with fellow developers, so I created DevMate! Find, connect, code, and meet with developers worldwide. 😉 cc. @_buildspace @_nightsweekends
7
6
333
Day 25 of #365DaysofCode ✅ quick sort, merge sort and recursion ✅ Recursion ques @LeetCode ✅ started building the project - DevMate @_buildspace @_nightsweekends #Connect #letsconnect #BuildInPublic #connect #webdevelopment #web
3
5
182
I love meeting new people and collaborating with fellow developers, so I created DevMate! Find, connect, code, and meet with developers worldwide. 😉 cc. @_buildspace @_nightsweekends
6
1
19
680