Mark

An indie hacker, know more at

[Mark @mark_lovstudio]

@modikr tks very much~

[Mark @mark_lovstudio]

yeah we used tantivy to speed up our indexing process a lot

[Himanshu Kumar @codewithimanshu]

@GitHub_Daily Lovcode's indexing process reduced Claude Code retrieval time by approximately 60 percent.

[Mark @mark_lovstudio]

@GitHub_Daily lol, much tks to recommend my project !

[GitHubDaily @GitHub_Daily]

使用 Claude Code 辅助开发项目，时间一长，对话记录越来越多，想回顾之前某个功能是怎么实现的，或者查找某段代码的讨论记录，翻起来特别费劲。

可以试下 Lovcode 这个开源桌面工具，专门为 Claude Code 等 AI 编程工具提供配套管理功能。

提供全文搜索的对话历史查看器，能快速检索所有项目的聊天记录，还能管理斜杠命令、模型控制协议(MCP)服务器配置等功能。

GitHub：https://t.co/uK9TgEPKL2

除了基础的命令和配置管理，还支持可复用的技能模板、自动化触发钩子、自定义 AI 智能体，甚至提供了社区模板市场可以直接安装使用。

跨平台支持 macOS、Windows 和 Linux 系统，可以直接下载安装包使用，也支持从源码构建。

[Mark @mark_lovstudio]

@hongkongdoll @Moon1ightSt 专业写插件，快速开发~
#公众号：手工川

[Mark @mark_lovstudio]

english transcription at: https://t.co/YuzxmDzYHW

chinese one at:
https://t.co/n4Zk720sVa

[Mark @mark_lovstudio]

Thanks for @lexfridman's AWESOME work!

I built a workflow for easier distribution to Chinese users via @AssemblyAI, published at https://t.co/sdwOT1XIGM, with codes at https://t.co/rMcQwJz9Wu.

I’m unsure if you’d like me to continue translating your videos. Is this okay?

[Mark @mark_lovstudio]

Thanks for @lexfridman's AWESOME work!

I built a workflow for easier distribution to Chinese users via @AssemblyAI, published at https://t.co/sdwOT1XIGM, with codes at https://t.co/rMcQwJz9Wu.

I’m unsure if you’d like me to continue translating your videos. Is this okay?

[Lex Fridman @lexfridman]

Here's my 5-hour conversation with @dylan522p and @natolambert on DeepSeek, China, OpenAI, NVIDIA, xAI, Google, Anthropic, Meta, Microsoft, TSMC, Stargate, megacluster buildouts, RL, reasoning, and a lot of other topics at the cutting edge of AI. This is was a mind-blowing, super-technical, and fun conversation.

Yes, we discuss r1 and o3-mini, but more importantly we look into the future of technology, geopolitics, and humanity in a world that stands on the precipice of a global AI revolution.

The first 4 hours are here on X (4 hours is current limit), and the full 5 hours are up everywhere else. Links in comment.

Timestamps:
0:00 - Introduction
3:33 - DeepSeek-R1 and DeepSeek-V3
25:07 - Low cost of training
51:25 - DeepSeek compute cluster
58:57 - Export controls on GPUs to China
1:09:16 - AGI timeline
1:18:41 - China's manufacturing capacity
1:26:36 - Cold war with China
1:31:05 - TSMC and Taiwan
1:54:44 - Best GPUs for AI
2:09:36 - Why DeepSeek is so cheap
2:22:55 - Espionage
2:31:57 - Censorship
2:44:52 - Andrej Karpathy and magic of RL
2:55:23 - OpenAI o3-mini vs DeepSeek r1
3:14:31 - NVIDIA
3:18:58 - GPU smuggling
3:25:36 - DeepSeek training on OpenAI data
3:36:04 - AI megaclusters
4:11:26 - Who wins the race to AGI?
4:21:39 - AI agents
4:30:21 - Programming and AI
4:37:49 - Open source
4:47:01 - Stargate
4:54:30 - Future of AI

[Mark @mark_lovstudio]

@AntWilson @supabase Really impressive!

[Ant from Supabase @AntWilson]

Happy 5th Birthday Supabase

In the last 5 years we’ve made countless mistakes - too many to list - so instead I’m going to write out some of the big things I think we got really right and helped turn @supabase into the product, company, and open source community it is today.

Finding the right co-founders

Deciding who to work with is the very first decision we got majorly right. From my perspective there’s very few people in the tech world who work as hard as @kiwicopple does. As a pair we don’t labor on decision making, I think this comes from each being willing to compromise our position because we believe in the competency of the other, and it turns out this scales further than just co-founders. Our early hires all behave in this way. We trust each other’s ability to execute and own large parts of the business. @Rorstro for example runs everything revenue related, and @everConfusedGuy runs product and engineering. Both quit their previous startups to work on Supabase and the founder mindset that they both possess means they take on responsibility by the truck load without needing to be asked. There are ex-founders in many key positions in Supabase, and it shows in the flexibility of the organization. The team have effortlessly adapted to large and ongoing changes in the way software is being built, both inside of large corporations and by fast growing one-person startups.

Viewing Support as part of the product

I think the first “Head of” we appointed in year 1 was a Head of Support. We always saw our ability to offer a truly world class support experience as a major advantage we would have over competing services. We even went as far as writing “Your number one responsibility is front line customer support” at the top of everyone’s contract, regardless of role. And we still do that today. We don’t care if you’re CFO or designing UI components, if you see a customer experiencing an issue then it’s your duty to assist. Support will never be an afterthought at Supabase.

Once we grew past just a handful of team members our remote first nature enabled us to fairly quickly offer 24/7/365 support. Much earlier than a typical organisation of our size would usually be able to do so. Now, having devs and support engineers across more than 30 countries means there’s never a moment in which we won’t be around to assist with potential issues.

Choosing Remote first

As it turns out, round the clock support is just one of many benefits of having a remote and globally distributed team. We were part of a small cohort of tech companies founded around the start of 2020 that due to The Great Remote Revelation (something which I feel has since been mostly forgotten) were not expected to move to San Francisco to unlock funding. We always felt that all the most important work is done hands-on-keyboard, and so far - this continues to be true.

Being fully globally remote and with no geo-adjustment on salaries means that we can hire the very best person for the job regardless of where in the world they happen to live and without needing to trade off on our requirement of being an incredible human being who is a pleasure to work with every single day. This has been great for culture, and fantastic for the quality of life for the folks who work at Supabase.

Making Everyone an Owner

A couple of days ago I got asked by a technology leader of a much larger, more established business, “if all of your staff work remotely, how do you know they’re working and not playing video games all day?”. My answer was one of motivation, and one of the ways we choose to motivate folks is through meaningful company ownership with a 10 year exercise window. A challenge here however is that folks in different geographies value stock in very different ways. Those familiar with Silicon Valley startups typically want as much stock as they can get. Some others, who may not be that familiar with the dynamics of fast growing startups would sooner choose an additional $1 of cash over $50 of equity. We educate new joiners as much as possible on the potential upside of Supabase stock and are as transparent as possible with everyone within the company on the dynamics of fundraising and the impact it has on their ownership. We’ve found this does indeed translate into some incredible outcomes (both for the business and for Supabase team members individually), and is something we’ll continue to invest heavily in.

Building the Platform from Day 1

A lot of open source software companies start self-hosted and offer services before later expanding into a hosted cloud offering (think MongoDB). At Supabase we decided to offer a managed service from day 1. It wasn’t as obvious a decision as it might seem today. Managing databases is a capital intensive business requiring a large team of experts, constant security and abuse management, and an extremely capable support team. Something we had to build from scratch. But boy are we glad we did it this way around. Platform considerations inform product more than you might think, and scale is built naturally into everything we do. It was also a more natural path to monetization, since the part of running a database people typically hate the most is configuration and infrastructure management - folks scaling fast readily hand that part off.

Don’t Build, Don’t Buy

Being fully open source wasn’t so much a decision for us as a core belief that this is objectively the best way to build sustainable software. There was one decision we made early on related to OSS however that helped us on our way. Whenever faced with demand for a new product or feature, we start by scouring the earth to find an existing open source tool and associated community we can adopt and embrace. As a good example, all the Go server code for Supabase Auth was originally written and maintained by @Netlify. We backed @andrewkane's pg_vector project early on before it went on to revolutionize the vector storage space, and we also heavily invest in the long term sustainability of the @postgrest_org project which powers our Auto-generated APIs. There are times of course that suitable tools don’t exist (perhaps they’re written in Java) and we’re forced to build or buy, but default mode is to Adopt and Support.

Taking the right partners

During the ZIRP (zero interest rate phenomenon area) period and with clear product market fit and traction we had the rare luxury of lots of potential suitors when it came to raising money. We started by going round all the smartest techies we knew - folks like @swyx, @biilmann, @Chr_Bach, and @rauchg and asked them to invest. These folks continue to give incredible advice and in the early days helped bootstrap our product marketing efforts across social media. We took funding from @ycombinator which was a natural launch pad for a business like ours, and went on to build a close relationship with @DavidCahn6 and @carynm650 of Coatue. Caryn, having led comms at Meta across their most explosive eras of growth, I feel will never truly be impressed by our growth numbers - and she’s right: capturing developer mindshare is not enough, we’re seeing now that AI powered developer tools are actually expanding (read: exploding) the TAM of DevTools - and the companies able to position themselves effectively will grow orders of magnitude larger than any previous generation company ever could.

We Rode the Big Waves

In the five years we’ve been building Supabase, there have been about as many major tech movements. Each time we managed to strike early addressing the emerging usecase without straying from our core competencies. Make no doubt that Supabase is a database company, and it turns out just about everyone/everything needs a database (Metaverse devices, NFT marketplaces, RAG systems, AI agents). Choosing Postgres for it’s exensibility was the best decision we made in our 5 years, and we now have an entire team of Postgres Hackers dedicated to making Postgres better for everyone. If I’m writing this post again in 5, 10, or 20 years - it will be because we doubled down on Postgres every time.

I’m eternally grateful for everyone in the team and community who has helped us over the last 5 years, and to those who continue to support us every single day. I can’t wait for what comes next.

[Mark @mark_lovstudio]

@zzzwco why

[Mark @mark_lovstudio]

(3/3) From my first knockoff "fruit machine" (LOL, did Jobs predict Fruit Ninja? 🍉🔪) to finally splurging on that rose gold iPhone 6s in college (judge me, I dare ya 😎), Apple's been my ride-or-die.

[Mark @mark_lovstudio]

（1/x) 10/06/2024: How Jobs Rocked My World 🍎💥

[Mark @mark_lovstudio]

(2/x) Yo peeps! 👋 Just had a mega Jobs-piphany on his 13th death anniversary. Buckle up for this rollercoaster! 🎢

[Mark @mark_lovstudio]

@GPTDAOCN @memdotai mem it

[Mark @mark_lovstudio]

@GPTDAOCN @memdotai mem it

[Mark @mark_lovstudio]

@elonmusk @memdotai mem it

[Mark @mark_lovstudio]

@elonmusk @memdotai meeeeem it

[Mark @mark_lovstudio]

@RyanMfer 【4】
太难挖掘到了，但不应放弃这样一种挖掘的态度和坚持

不过可以先通过各种方式活下来，再等待那个痛点的降临，然后一击必中！

[Mark @mark_lovstudio]

@RyanMfer 【3】
很可惜对于绝大多数初期创业者，1、2大致是具备的，3和4往往互不相容

2B的创业者大多数会抛弃1，最后找到 2 + 3 + 4 的交集

2C的创业者，则几乎是 1+ 2 + 3，4 很难，除非能像爱彼迎一样真地挖掘到市场上暂时还空缺的生态位

[Mark @mark_lovstudio]

@RyanMfer 【2】
尽量找结合点，也就是：
1. 自己感兴趣
2. 自己擅长（1可以转化成2）
3. 社会需要（有尽可能多的人愿意买尽可能多的单）
4. 竞争对手少

1、2、3的结合点，就是一门好生意，如果4也成立，那就是顶好的生意

如果一个创业者能找到1、2、3、4的交集，那大概率能成