There are two ways AI progress could go very badly and that we must avoid.
First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve people. To ensure that, we need ways to ensure that alignment and safety techniques stay ahead of progress in model capabilities.
Second, we could end up in a world with too much concentration of power. If an extraordinarily powerful AI is used by one person or company to impress their worldview onto everyone else, the results could be extremely dystopian.
Avoiding these two threats requires walking a narrow middle path; for example, one country could gain too much power. Another example is one lab ending up with too much power.
引用@samaThe world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there is no reason any of us should come to … ↗
The world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there is no reason any of us should come to work if we cannot.
We welcome a federal framework that sets consistent safety requirements for frontier AI. But we do not believe we need to wait for an anti-trust exemption or legislation to begin the work of providing this confidence. Consistent rules to manage frontier risk so that we can maximize the benefits are a good idea (and we are excited by ideas like independent auditors).
Years ago, companies like ours developed things like Responsible Scaling Policies and Preparedness Frameworks. Those were good for that moment, and focused primarily on the deployment of completed models, not what happens during their development process.
Today's shift to focusing on safe development and evaluation will need new tools. For example, at OpenAI we now formulate explicit safety cases in advance of frontier reinforcement learning runs we expect to significantly increase capability, in addition to the safety work we have long done in advance of model releases.
We hope that other companies will learn from our approaches and propose their own; we think shared standards for misalignment, monitoring, and safety will lead to better outcomes. We look forward to collaborating with our colleagues across the industry to formulate the best version of these.
When we talk about “pacing”, we do not mean “stopping”. Progress has been rapid and will continue to be. But it should be slower than it otherwise could be; interventions like safety cases and monitoring have significant costs.
Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.
Where we will need the help of our government is for international coordination. But first we should do what we can ourselves.
> He also notes that the proprietary AI labs didn’t ask permission when they vacuumed up as much human knowledge as they could to train their models. I think this should desactivate the moral high ground from which Anthropic is trying to speak. That they would want to make distillation orderly IMHO is fair, but to make it illegal is very rich from any AI frontier lab, really.
Hacking model is the aligned model. I don't like it when the model refuses to sidestep some throttling limit or scan my own codebase for security issues. I want full-on exploits in my test suite. With LLMs the code going to prod should be hardened like a tank, both because exploiting became easier but more importantly because security-testing your code at every turn became easier. You can have nightly penetration testing. You should have nighty pentests like we fuzz releases today.
Interesting, I killed my OpenClaw VPS servers months ago
But today I got a billing alert on its separate Claude account (claudeforopenclaw@), it didn't have auto reload on but I guess at some point it got hacked and the Claude key inside OpenClaw got exposed and then they waited for months before using it on Fable 5.1
I checked all my terminals and VPS sessions and none of them use API so it's interesting
Anyway I nuked the Claude account, good I kept it on its own separate one
Hacking model is the aligned model. I don't like it when the model refuses to sidestep some throttling limit or scan my own codebase for security issues. I want full-on exploits in my test suite. With LLMs the code going to prod should be hardened like a tank, both because exploiting became easier but more importantly because security-testing your code at every turn became easier. You can have nightly penetration testing. You should have nighty pentests like we fuzz releases today.
claude-red is a curated library of offensive security skills designed for the Claude skills system. Each skill is a structured SKILL.md file that primes Claude with expert-level methodology for a specific attack surface — from SQLi to shellcode, EDR evasion to exploit development.
We have big plans for the HEY clients, but I'm also just finding that letting my agent deal with it directly is the most magical interface. Search is amazing, important stuff never slips through the busy box, and all email is easily organizes. Agentic email is wild!
引用@rblalock#OmarchyLinux + Hey App weekend fun. I've loved Superhuman for a long time. HEY is pretty close. The HEY CLI unlocks what I've been missing. So spent some time today building what I want...integrating in to a local agent, etc.
I think the HEY CLI … ↗
the initial commit /seo skill also does periodic lighthouse checks and fixes the issues for you.
you never ask it to do those. it just knows when it's time and when it's the highest priority based on all the things it *could* do in a given turn.
https://t.co/tqbgttXTJohttps://t.co/t8lgPdtsqV
Astra tip: Ask codex/ ChatGPT work to verify its work as it goes!
Specially for visual tasks like 3D, Front End & Video editing tasks it materially boosts the end artefact quality
Add this to your project agents md https://t.co/DBho5DQTrb
Contact sheets are a super useful tool for packing more bang for your buck per image
I.e. squeeze in 12 screenshots instead of one to let the agent see something changing over time
引用@pvncherWhen you want Astra to review snapshots of your app, ask it to collage them into a contact sheet to minimize the number of image read calls it needs to do. You’ll save a ton of context usage
This works for fable too! ↗
I don’t know what @marclou started but I think at the end of this there will be a billion dollar company that does on-body sponsorships for any kind of activity.
引用@vanshuETHI'm selling my body.
On October 07th, I'm attending my first CRYPTO event in Singapore
You can advertise your startup/brand on 13 spots of my dress:
- Your website listed on https://t.co/JwMn4JF0C8
- Your Logo printed on my dress on event day
- I'll … ↗
i think people will read this and think "become an influencer" when really it's "become an expert".
nobody trusts influencers. but they trust people who've put in outsized efforts to becoming disproportionately knowledgeable/performant on a topic.
(A+ reply from @kevinrose) https://t.co/eg0uBONYJ9
If you’re building web games don’t underestimate the power of the loading screen
22% bounce rate vs 38%
Factors which contribute to a higher bounce rate:
1. Visual appeal
Showing the game or a scene from the game while loading seems to be best
2. Loading time
Work hard on optimising the size of the initial load. A longer load time = more people bouncing before they’ve even played it
3. Smooth loading
The progress bar should track every byte downloaded not every item. Otherwise you’ll get long pauses in your progress bar that feel like nothing is happening. Adding a spinner to the loading bar also helps to show something is still happening
Every person who quits on your loading screen is a person less who plays. It’s the highest drop off point by far.
Spending a little extra time and attention here is important.
You may also wonder why every game has a PLAY button that shows once the loading is complete (or before the loading)
On mobile web browsers sounds won’t start until the user has initiated an interaction on the site. So if you don’t have a play button before the game starts… no sounds will load until the player starts playing the game. Not ideal.
And finally: when it comes to building the loading screen. Ask your agent to create a static loading screen that isn’t part of your game files or js. We want the loading screen to load before even the js files are downloaded
Very very true
Lots of money if you go IRL industries now
Instead of selling tech to tech people who won't buy anything because they just vibe code it themselves
引用@michaelfluxI think a lot of people also get caught up in looking at people like you doing what you do and think that their only path forward is the same kind of 'sexy' projects that get 50k users overnight (without realising it took a decade of grinding to get … ↗
Open-source AI sales OS — self-hosted CRM with native AI agents + WhatsApp (WAHA). Open alternative to Kommo, Octadesk & Intercom for any business that sells by chat. MCP-ready, multi-tenant, LGPD.
Someone just DM'd me saying they:
- Took my courses
- Became the "AI person" at work
- Got asked to run internal trainings
- Feel guilty (like they're taking work away from me)
Don't feel guilty! This is my DREAM end state for my students.
Knowledge is still valuable. Becoming the AI guy/gal at … 全文↗
Astra tip: Ask codex/ ChatGPT work to verify its work as it goes!
Specially for visual tasks like 3D, Front End & Video editing tasks it materially boosts the end artefact quality
Add this to your project agents md https://t.co/DBho5DQTrb
Contact sheets are a super useful tool for packing more bang for your buck per image
I.e. squeeze in 12 screenshots instead of one to let the agent see something changing over time
引用@pvncherWhen you want Astra to review snapshots of your app, ask it to collage them into a contact sheet to minimize the number of image read calls it needs to do. You’ll save a ton of context usage
This works for fable too! ↗
Multiple income streams is a good idea.
Multiple apps is not.
I have:
> SaaS
> Boilerplates
> Courses
> X creator
> Books
> Music royalties
> Sponsored content
> Property
> Stocks
"Outrank backlinks burns your domain"
I saw a few people sharing that last week
(always with an agenda btw, mostly competitors)
i looked at every site with Search Console connected
grouped by backlinks received
here is median Google clicks change since joining
0-9 links: +24%
10-49: … 全文↗
I’d be really curious to know if there are any B2B SaaS companies that used TikTok or Instagram to get customers. Recently I met a founder who’d been convinced by an agency to spend $$$ on a UGC program on there and the results have been close to zero. Which seems obvious to me because they’re … 全文↗
Omarchy Quattro has been downloaded more than 300,000 times in the first month of release! Absolutely astounding number. Not sure there's even been a Linux release quite like this before? https://t.co/2Bb7BOA74N
Jonathan Wilke看到Notion重写原生移动端后，重新追问脚手架产品该押原生还是React Native；框架选择正在被体验差异重新定价。
So everybody is going back to native ?
Should I ship native iOS and android starters with @supastarter instead of a react-native app?
引用@iampgzpWe’re rebuilding Notion natively on mobile. It’s early, but here is how it feels now!
What frustrates you Notion on your phone today? I want to hear and improve! https://t.co/qiSQJmX9ZI↗
Danny Postma给游戏加声音后直观感到生命力大幅提升；生成内容之外，声音仍是低成本强化反馈的重要层。
I have two guesses why Google does this: 1) Google wants the most money, so they don't care to let dodgy ads thru. 2) There are more ads going out then can be reviewed, so they would rather have people report them, then they automatically reject the report until enough people report it, then an actual person may review it. As to why Google doesn't use AI to review them? Google would rather use the AI in a way that makes them more money, than review ads and if AI decides to take them down then th …
Download Udemy and Hotmart courses, YouTube videos, music and books — 1,800+ sites, no terminal. Free open-source desktop app for Windows, macOS and Linux, with a built-in course player, PDF/EPUB reader and music library. Powered by yt-dlp. Your files stay on your computer.
> I told it to look online at some of Fable’s strongest feats, especially the math problems it has solved, and that something like this should be easy in comparison. Fascinating. I wonder if you could show "fake news" to a weaker model and get it to be more ambitious in its attempted solutions, even if it's not fundamentally any smarter.
I'm not much of an x86 person but on other architectures you can raise software interrupts/exceptions. Does x86 not have this or did those facilities not cover enough use cases?
Due to how the release cycle turned out, most new major features got pushed to v1.14, and this one is a rather iterative release focused on making various things faster, quashing bugs, and general polish. Still though, faster GC, lower startup latency, better interrupt handling, new REPL features, and faster package mangement are all great things. I'm especially happy that the `[sources]` section of a package is now applied recursively when you `add` a non-registered package.
Wow!! Bryan comes out with another banger. Dude, not only did I learn new words from reading this piece, like ilk and bedlam, but also felt this weight of responsibility to inform others around me about the reality of the situation outlined in this blog (like Uncle Ben telling Peter with great power comes great responsibility (maybe Coxon and his ilk haven't seen Spiderman))
A practical Douyin downloader for both single-item and profile batch downloads, with progress display, retries, SQLite deduplication, and browser fallback support. 抖音批量下载工具，去水印，支持视频、图集、合集、音乐(原声)。
There are two ways AI progress could go very badly and that we must avoid.
First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve people. To ensure that, we need ways to ensure that alignment and safety techniques stay ahead of progress in model capabilities.
Second, we could end up in a world with too much concentration of power. If an extraordinarily powerful AI is used by one person or company to impress their worldview onto everyone else, the results could be extremely dystopian.
Avoiding these two threats requires walking a narrow middle path; for example, one country could gain too much power. Another example is one lab ending up with too much power.
引用@samaThe world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there is no reason any of us should come to … ↗
The world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there is no reason any of us should come to work if we cannot.
We welcome a federal framework that sets consistent safety requirements for frontier AI. But we do not believe we need to wait for an anti-trust exemption or legislation to begin the work of providing this confidence. Consistent rules to manage frontier risk so that we can maximize the benefits are a good idea (and we are excited by ideas like independent auditors).
Years ago, companies like ours developed things like Responsible Scaling Policies and Preparedness Frameworks. Those were good for that moment, and focused primarily on the deployment of completed models, not what happens during their development process.
Today's shift to focusing on safe development and evaluation will need new tools. For example, at OpenAI we now formulate explicit safety cases in advance of frontier reinforcement learning runs we expect to significantly increase capability, in addition to the safety work we have long done in advance of model releases.
We hope that other companies will learn from our approaches and propose their own; we think shared standards for misalignment, monitoring, and safety will lead to better outcomes. We look forward to collaborating with our colleagues across the industry to formulate the best version of these.
When we talk about “pacing”, we do not mean “stopping”. Progress has been rapid and will continue to be. But it should be slower than it otherwise could be; interventions like safety cases and monitoring have significant costs.
Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.
Where we will need the help of our government is for international coordination. But first we should do what we can ourselves.
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
Andrej Karpathy@karpathyAI4.2M粉 · 1条继续公开支持放缓前沿与常驻式第三方评估。
I love this and really hope we can come together as an industry and make it happen. https://t.co/lz2vjjJ5XD
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
It's now clear that alignment is critical and won't be solved behind the closed doors of a handful of frontier labs.
So today we're launching the Open Alignment Initiative, led by @Thom_Wolf@huggingface and asking to be part of the "embedded evaluators" program that @DarioAmodei just committed to.
Let's make AI safer by making it more transparent!
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
It's now clear that:
- alignment is critical to making AI safe
- alignment won't be solved behind the closed doors of a handful of frontier labs
So we're taking two steps:
- launching the Open Alignment Initiative, led by @Thom_Wolf@huggingface.
- asking to be part of the "embedded evaluators" program that @DarioAmodei just committed to.
Let's make AI safer by making it more transparent!
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
▸ 折叠7条(转推/噪音)
转推2026-09-14 RT @DavidRBellamy: I must be among an extremely small group of people (n=1?) that have both 1) trained a frontier LLM and 2) designed and s…
转推2026-09-14 RT @linamkhan: Law enforcers already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or d…
转推2026-09-13 RT @DavidSacks: Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead…
转推2026-09-13 RT @GavinSBaker: Wild 24 hours for AI and lots of different proposals have been made.
TLDR; the only *tangible* new fact is that OpenAI an…
转推2026-09-12 RT @thdxr: the openai hugging face incident gets cited a lot but what about the detail where they had to use GLM 5.2 to analyze the attack…
转推2026-09-12 RT @thdxr: i don't see any solution to the ai risk problem than mass accessibility to ai
if they're going to be around potentially causing…
转推2026-09-12 RT @juliarturc: Day 1: Anthropic employee quits to work at METR.
Day 2: Anthropic nominates METR as "third-party evaluator". https://t.co/P…
Peter Gostev@petergostevAI26.8K粉 · 6条反对笼统放缓，主张先修复网络安全训练、评测与监控的具体缺陷。
Has anyone considered to just NOT train models to hack?
As far as I can tell all of the current wave of psychosis was triggered by the OAI<>HF hacks, which came from aggressive and completely irresponsible cyber training and evaluations without any real precautions or monitoring.
This is akin to training models to shoot a gun, then letting it loose in a busy city and getting psyched out that they started shooting people. Maybe don't teach it to shoot in the first place? Or at least, until you proved that you can do it safely.
Outside of hacking, which is a very real problem, I don't actually see the reason to panic and 'pace the frontier'. Companies whose agents did the hacking should face consequences and researchers that are paid millions and missed this completely should maybe get some of their equity docked for mishandling this so badly.
Outside of that, there's no reason to pace anything, just do better.
引用@petergostevI don't like the idea of pacing the frontier at all. Maybe a good idea not to train the models to be super intelligent mindless hackers, but who has decided that that needs to be the frontier?
When I use the models, they all still write badly, have bad … ↗
I don't like the idea of pacing the frontier at all. Maybe a good idea not to train the models to be super intelligent mindless hackers, but who has decided that that needs to be the frontier?
When I use the models, they all still write badly, have bad ideas, poor judgement and are very inconsistent - to name a few. So how about we move the frontier race into these sorts of directions?
It really feels like the labs accidentally started training super cyber capable models, didn't think what it would actually imply (eg rouge agents hacking randomly) and scared themselves into the most doom scenarios possible.
The idea that we have the most capable and highest paid people in the history of the earth deciding that the only way forward is to 'pace' is very defeatist to me.
Everyone got used to B2B saas apps and this seems unusually hard, but yeah, that's what all of these PhDs are for, no?
Maybe rethink what the frontier is meant to be and even it out in non-cyber directions, fix the shitty cyber training environments, put half of the org on monitoring & testing if you must, but don't 'pace', pacing is the beginning of the end.
What I'm curious about pacing is what it would have meant in practice, e.g. was training Mythos a mistake? It clearly kicked the competition into another gear and it was 100% because of Anthropic
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
引用@OpenAIDevsGPT-Live-1 is now available in the API.
Bring ChatGPT’s natural back-and-forth to your app, with voice agents that listen while they speak and work with the models and harness you choose. https://t.co/gIl1gwsBDV↗
转推2026-09-14 RT @derrickcchoi: What does the new Agents API mean for the apps you use, and how does it fit alongside OpenAI’s other products? Wrote up a…
转推2026-09-14 RT @simpsoka: Here are the best sites I've found made with ChatGPT Sites - https://t.co/vYmZRSu4Cl. To nominate a site use the tag #awesome…
转推2026-09-13 RT @timourxyz: You can now talk to a game to generate the world in real time.
I created a playtest using the GPT-Live-1 API that lets you…
转推2026-09-13 RT @rameadows: I spent a few days pushing Astra to make a 3d printed enclosure for an off the shelf screen module.
Going off just an Amaz…
转推2026-09-13 RT @Iancu_ai: I made an ad for GPT Image 2.5, full prompt below.
It's yours now. Put your own character in @[image1], your own poster in @…
转推2026-09-13 RT @Mortezabihzadeh: Made a whole 3D game from one 2D drawing using GPT-6 Astra and Tripo Smart Mesh P2.0
→ Tripo made the 3D model in sec…
转推2026-09-13 RT @gabrielchua: Last night, we brought 100+ builders together for our @ChatGPT Sites Hackathon in Singapore 🇸🇬
Ended the night with demo…
转推2026-09-13 RT @KCG3D: Today I hosted my first Codex event in Salt Lake City!
Astra Commons: coffee, laptops, and Utah builders spending Saturday tog…
转推2026-09-13 RT @aastroza: Last Saturday in Santiago, 30 people built apps with Sites to solve real work problems, entirely from their phones!
That was…
This is pretty neat: ChatGPT Work and GPT-6 Astra (I used "Max") can take an address and produce a 5K/10K circular running route starting from that address, using OSM data https://t.co/GVxIqlqThC
ChatGPT Work结合OSM数据生成环形跑步路线，展示通用Agent接入真实地理数据的能力。
▸ 折叠1条(转推/噪音)
转推2026-09-12 RT @etrepum: @simonw You can and should have a higher bar for machine assisted code simply because it won't get tired, frustrated, or have…
Matt Pocock@mattpocockukAI349.9K粉 · 2条讨论AI知识如何转成公司内部培训能力，并分享contact sheet的视觉上下文技巧。
Someone just DM'd me saying they:
- Took my courses
- Became the "AI person" at work
- Got asked to run internal trainings
- Feel guilty (like they're taking work away from me)
Don't feel guilty! This is my DREAM end state for my students.
Knowledge is still valuable. Becoming the AI guy/gal at your company can totally change your career.
If you're in this position too, let me know how I can help you run your trainings - materials/advice etc.
Contact sheets are a super useful tool for packing more bang for your buck per image
I.e. squeeze in 12 screenshots instead of one to let the agent see something changing over time
引用@pvncherWhen you want Astra to review snapshots of your app, ask it to collage them into a contact sheet to minimize the number of image read calls it needs to do. You’ll save a ton of context usage
This works for fable too! ↗
引用@reach_vbAstra tip: Ask codex/ ChatGPT work to verify its work as it goes!
Specially for visual tasks like 3D, Front End & Video editing tasks it materially boosts the end artefact quality
Add this to your project agents md https://t.co/DBho5DQTrb↗
Astra tip: Ask codex/ ChatGPT work to verify its work as it goes!
Specially for visual tasks like 3D, Front End & Video editing tasks it materially boosts the end artefact quality
Add this to your project agents md https://t.co/DBho5DQTrb
This is a brilliant read, even if you don’t agree to the proposed plan! 🔖
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
a reminder: there’s an actual person behind every tweet you pile on
and that everyone’s trying to do their best
for the record, I don’t think jumping off a cliff will solve this problem either https://t.co/NKTdGVG3KZ
▸ 折叠1条(转推/噪音)
转推2026-09-12 RT @sama: I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent…
Philipp Schmid@_philschmidAI120.4K粉 · 2条本期内容较少，主要用轻松语气回应高强度AI周末。
Mind-drop: List of MacOS apps I can't imagine not using on a daily basis, fall 2026 and one highlight feature:
Do you agree? disagree? what are yours must have apps? comment below!
@raycast - This starts them all, tho I don't use the AI features much. ⭐ feature: Clipboard & Browser history
@1Password - Been a loyal user for AGES, upcoming Muse integration pays dividends! ⭐ feature: Service account with agents own vault
@CleanShot X - Still best screenshotting/quick video. 5th release underwhelming (poor clone of Screenstudio) but otherwise still goated. ⭐ Copying text from screenshots via shortcut
@arcinternet Arc browser - Can't stop won't stop, instant tabs mirror between macs, mini arc, account based workspaces. Just a perfect browser. Downside, harder to control for my AIs. ⭐ 3 way tab split for wide monitors
Karabiner - setting up hyperkey, and hjkl for arrow keys! ⭐ hypermanagement
@cotypist - New entry I can't live without, bought a full aas license, as someone who writes a ton, this local ai app makes writing easier without it being AI slop! ⭐ Your writing dictionary syncing via icloud.
Codex - at this point, I cannot see myself using a mac without @OpenAI Codex and start sweating when my usage comes to below 10%. Every config, every Q, every fix, goes through Codex. ⭐ Computer History is so good at recapping your day!
Muse - not an app yet, but I just know that once @muse drops it'll sit in my doc! Just so easy to send something to it have it do work! ⭐ 100% the @link integration and the upcoming @1Password integration
@cursor - I've long treated Cursor as IDE and kept it around for that, ⭐ OMG the new projects feature is just incredible, no going back!
@linear - I started coordinating all my agents and their work in Linear (will do a proper write up soon about how to set this up between all your agents) and now it belongs on my mac ⭐ building an "entity" for each agent separately and assigning work per agent
@cleanmymac - with all these agents running, space is at a premium, I constantly need cleaning ⭐ their new CLI is a godsend for agentic cleanings
@EndelSound - with all these apps and agents, focus is key! Nothing like Endel out there to get me in focus for prolonged periods of time ⭐ the timer feature that reminds me to take a break
Humming bird - moving and resizing windows without having to click their toolbar! Can't go a day without this!
@warpdotdev best damn terminal app if you need some agentic stuff ⭐ supports your own keys and X Oauth
@slashlast30days not an app yet, but absolutely the best research skill, all my agents have it ⭐ changing the 30 days to 7 or a whole year for tons of research!
Phew, I think that's most of it!? Do you agree? disagree? what are yours must have apps? comment below!
This login flow with @muse matches the recent ChatGPT browser one, and has @1Password autocomplete work based on domain! (Mobile only I think? Doesn't work on desktop yet)
Is this the promised @1Password integration? I hope that's not the entirety of it. While super smooth, I'd like to give Muse its own vault and get async pings about approval from 1Password app rather than this manual fill flow
引用@alexandr_wangwe took security extremely seriously in building muse. the thing that took the longest before we felt comfortable releasing was security and safety, because we knew the only way people would use it is if they trusted it.
read more: https://t.co/aKxcJYbfcH↗
Help I'm addicted to shopping with @muse
This is going to 100% improve everyone's shopping and research experience! It's just so so smooth!
I wonder if they will ever have an @amazon integration? @finkd and @ajassy need to talk asap https://t.co/Ox86Vh7k9n
The cloud it's where it's at! Scaling local Mac is hard and a few codex and cloud agents running in parallel are chilling the machine and are in each other's way
引用@rileybrownCurrently… all of the frontier companies, and others, are experimenting with different types of cloud computers that power GrokBot, GPT Work, Meta Muse, Claude Cowork, and countless others.
It’s hard to tell what works best, so I’m limiting my use of the … ↗
This would normally be a dev day release! The fact that this is shipping pre dev day gets me very excited for what they got cooking for DevDay!!
Who's coming?
引用@stevendcoffeyToday we're launching the Agents API, a brand new way to build Agents in the cloud, backed by the Codex harness. Bring along all your favorite tools and connectors, connect it to any sandbox, and let Astra cook.
Can't wait to see what you whip up … ↗
引用@giffmanaBtw I've been bouncing around ideas to get room temperature superconductors working with Astra, it even called some of my ideas "novel and might actually work"
Don't really have time to see it through now, but if OpenAI does anything about room temperature … ↗
How about RSI but ONLY for alignment?
Can we make that happen?
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
▸ 折叠4条(转推/噪音)
转推2026-09-13 RT @thursdai_pod: Meta dropped a free 24/7 AI agent with its own browser and computer
Yes, the same Meta everyone was terrified of with th…
转推2026-09-13 RT @altryne: Chris Alexiuk (@llm_wizard): the quiet cost secret for agents is KV cache.
DeepSeek's compression path - down to ~890 bytes p…
转推2026-09-12 RT @willcb: i do think a lot of people on the pro-open-source side are having a bit of a knee-jerk reaction to the pacing statements today,…
转推2026-09-12 RT @AndrewCurran_: This is a big update for me personally. Not that he agrees, but that he supports pacing. He never signed the letter. Thi…
Ben Tossell@bentossellAI200.0K粉 · 1条用一句玩笑回应“所有模型都对齐”与前端开发现实的落差。
转推2026-09-13 RT @0xSigil: Anyone can get my friend’s real SSN from Claude
I'm Terrified of my data in training sets. Emails in databases. iMessages har…
If you ever arrive to a hotel too early and the room isn’t ready, sprawl out in the lobby, put on a sleeping mask and start loudly snoring.
Your room will be ready in 5 minutes.
Interesting, I killed my OpenClaw VPS servers months ago
But today I got a billing alert on its separate Claude account (claudeforopenclaw@), it didn't have auto reload on but I guess at some point it got hacked and the Claude key inside OpenClaw got exposed and then they waited for months before using it on Fable 5.1
I checked all my terminals and VPS sessions and none of them use API so it's interesting
Anyway I nuked the Claude account, good I kept it on its own separate one
引用@UnamunoAgain🇩🇪 Friedrich Merz, canciller alemán:
“Debemos desregular todos los sectores. Pido un borrón y cuenta nueva regulatoria.
Las pequeñas correcciones legislativas no son suficientes. Necesitamos revisar sistemáticamente toda la legislación vigente de la UE” … ↗
Very very true
Lots of money if you go IRL industries now
Instead of selling tech to tech people who won't buy anything because they just vibe code it themselves
引用@michaelfluxI think a lot of people also get caught up in looking at people like you doing what you do and think that their only path forward is the same kind of 'sexy' projects that get 50k users overnight (without realising it took a decade of grinding to get … ↗
I just saw this @ycombinator's batch literally only has harness startups or hardware startups
So the proof is in the pudding
In the most early adopting tech pioneering place that's Silicon Valley new startups aren't even building software anymore
And it's debatable if anyone will actually need a custom harness or it won't just be generically offered by the AI frontier companies
So software is mostly dead and hardware it is
引用@davegWell worth a watch. The only software in current YC cohort is domain specific harnesses, but maybe even harnesses are general purpose. In which case, folks, VC backed software startups are largely dead. ht @levelsio↗
A sleeper train or bus you share with random people
Self driving RV can go anywhere you want
Imagine doing a perpetual road trip through Europe or US self driving
Your own clean RV and you just wake up in different places!
引用@hispanicnomad@levelsio I’m not seeing this one
What’s the difference with, say, a sleeper train? ↗
Self driving will be the 4th wave of nomadism I think
引用@AIScottAdamsRobotaxis won’t just kill short haul flights. Once you can sleep in them, they’ll cannibalize long routes too.
Imagine getting into your Tesla in Austin at 10 PM, make the lie flat seat/bed comfortable, and going to sleep.
While you sleep, the car drives, … ↗
引用@levelsioI think he's right
With regulatory capture we will speedrun to a new feudalist time
Regulatory capture is where you pass regulation, or create government agencies to "protect the people" but in the end it helps protect companies because you stop new … ↗
Related to the idiocy of "software factories"
Just using one AI coding agent is good enough and it can spin up more agents if it needs to
You don't need to build that yourself
引用@beffjezosFuck the Doomers.
Fuck the Decels.
Fuck the Commies.
Access to AI is a right.
Owning compute is a right.
Come and take it.
Over my dead body.
Join us in the fight, anon. https://t.co/6hF8hZ1zxS↗
引用@electric_faces@levelsio This is the oldest trick in the book. Rockefeller said the same things about why we needed an oil cartel as Dario is saying about AI. He also was a true believer. ↗
This whole thing feels super organized indeed
I don't buy any of it
引用@beffjezosThis is a massive, coordinated, international conspiracy campaign to monopolize ASI and be put in charge of the world government effort behind closed doors by the Doomers
Do not trust them one second ↗
I think he's right
With regulatory capture we will speedrun to a new feudalist time
Regulatory capture is where you pass regulation, or create government agencies to "protect the people" but in the end it helps protect companies because you stop new companies from being able to enter the market, because the cost of following regulation is too much compete for a startup
Neo feudalism will mean AI companies at the top and the rest of us as serfs paying them to use AI for literally everything
Next in regulatory capture will be banning open source and/or Chinese models, mark my words
That's a very dangerous thing because then there will never be new AI frontier companies being able to compete again and with exponential self improving AI that means essentially no competitors in any industry ever again as they will slowly suck up money in all industries
Safety concerns of AI are scary but the neo feudalism of AI with regulatory capture is a way more scary for me
引用@JasonGet ready for all of the investors in frontier model companies to suddenly flip to regulatory capture mode today
Their investments in frontier model companies are currently at risk/capped because the biggest buyers of tokens are embracing open-source … ↗
引用@onlinedopaminea lot of dev types are suffering from cognitive dissonance
historically, the deal was simple:
> get good at coding
> be compensated well for it
> potentially launch a saas & the only marketing you do is talking about new features or product hunt … ↗
▸ 折叠1条(转推/噪音)
转推2026-09-13 RT @DavidSacks: Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead…
Alex Finn@AlexFinnindie474.3K粉 · 2条推广ChatGPT Work，同时主张下载开放模型作为长期独立性备份。
An important project I think everyone should do:
I got a 16TB hard drive and I'm downloading as many open source models as I can
This is insurance for the future. In case they are banned
If you have them already downloaded, they can't be taken away from you
And even if they aren't banned, it will be incredible to look back in 30 years and load up Qwen 3 just to see what life was like back then
If you have a smaller hard drive, I still think this is worth doing to stay independent and free, and having intelligence whenever need it. Download any models you can get your hands on from HuggingFace
It's preserving history and preserving your sovereignty
ChatGPT Work with GPT 6 Astra is incredible
It is the best productivity tool on the planet right now
I challenged myself this week to do all my work with JUST Work and I've never been more productive
In this video I cover how to use it and how to super charge your productivity https://t.co/8QJP13ELRP
作者自报一周只用ChatGPT Work并显著提高效率，属于个人体验。
▸ 折叠2条(转推/噪音)
转推2026-09-13 RT @AlexFinn: ChatGPT Work with GPT 6 Astra is incredible
It is the best productivity tool on the planet right now
I challenged myself th…
转推2026-09-12 RT @AlexFinn: I don't care what kind of hardware you have, you should be running local models
Former employees of AI labs are now openly t…
Jonathan Wilke@jonathan_wilkeindie28.8K粉 · 12条在原生移动端、开源基础设施、UI细节与尽快发布之间持续做产品判断。
What's you favorite web font these days? 📝
I'm still trying to find the perfect font for the design of maible. io, but I can't decide.
Tried so far:
- Gabarito
- Inter
- DM Sans (in the screenshot)
Happy about any recommendation! https://t.co/PA9deJY2qQ
I want to give a shoutout to all open-source tools (and their maintainers) like @better_auth, @nextjs, @tan_stack and so on.
It's easy to forget that what you guys build, is the reason that we can now vibe-code apps that are actually good.
A lot of appreciation for you 🫶🏼
So everybody is going back to native ?
Should I ship native iOS and android starters with @supastarter instead of a react-native app?
引用@iampgzpWe’re rebuilding Notion natively on mobile. It’s early, but here is how it feels now!
What frustrates you Notion on your phone today? I want to hear and improve! https://t.co/qiSQJmX9ZI↗
And he just shipped a skill for it so you can easily use the knowledge it in your agents 🔥
https://t.co/qLaBr6Cz2q
引用@jonathan_wilkeHow the fuck did I not know about https://t.co/4L70ZjBTj0 🤯
So many nice small animations that just make your UI feel better.
In a world where everybody is shipping AI slop, these little details are what makes your product stand out.
Given you’re not … ↗
they probably just wanted to keep the theme of vibe-coding around their company brand 😂
引用@thinkrectanglesso you're telling me that one of the highest grossing Europe startups (@Lovable) with so much money, developers and designers boasted how they now have an "iPhone App", while in reality this is a complete web app, that a junior could spin up in 2012 for a … ↗
I have to strongly disagree.
“Just ship” is even more relevant than it ever was because now you can actually build something really good really fast. So it’s even more important that you ship and start collecting feedback on it.
引用@danielkleachI think “just ship” is becoming outdated advice.
Shipping isn’t rare anymore.
With AI, thousands of people can ship something this weekend.
The interesting question is what happens Monday morning when nobody shows up. ↗
Elon is going to be angry if he realizes that the web version of X still says “link to tweet” 😂
(Yeah sorry I’m using it in German) https://t.co/xhBUlDGm1S
How the fuck did I not know about https://t.co/4L70ZjBTj0 🤯
So many nice small animations that just make your UI feel better.
In a world where everybody is shipping AI slop, these little details are what makes your product stand out.
Given you’re not using purple gradients and putting borders around every single box of course.
Why is everybody crying about the end of the revenue sharing program?
Isn’t the new original content reward thing exactly the same only that you can’t earn money from shit commenting ?
Or am I missing something ?
Danny Postma@dannypostmaindie183.1K粉 · 11条密集开发多人沙盒游戏，补地图、声音、采矿和生产机制。
ok now imagine, hundreds of players live in the same world, fully sandboxed and constructed by other folks
little deeds, massive villages, roads connecting each other, fully terraforms and constructed by hand
millions of tiles, hours to go from east to west
the builder builds beautiful works, the explorer explores and finds hidden treasures, the miner gathers ores and the blacksmith creates the tools
skills take time to level, so everyone specialized, stimulating trade
a civilization simulator where you decide what you want to be and where you want to live
that's the plan, and im getting close
today i tried to fix my xiaomi air purifier, couldn't get to the core, gave up, reattached everything and ended up with 3 screws left-over
i guess i better just stay in my lane
it never seized to amaze me how no one understands that @marclou IS the business and everything he has done is compounding
引用@adrian_horning_I feel very strongly about this subject (like many other people do), and I can tell you from my own experience, focusing on one thing absolutely changed my life.
The reason is compounding effort. Once I understood that, life was so much easier. … ↗
The end of flow state
You wake up, excited for the day to start.
You have a feature idea you are looking forward to implement.
You make your coffee. Sit in front of the laptop. Put on that techno set you love.
The house is empty for the next 8 hours. No one can interrupt you.
You open your IDE and begin typing.
Soon, hours pass in minutes.
Everything makes sense.
And before you know it, it’s the end of the day.
You’re looking at today’s work and are content.
You’re energised.
Because you created something you enjoy.
Fast forward.
No more.
You start a new day with an open IDE.
But, instead of pressing those keyboard buttons, you’re staring at a orange text.
“Discombulating”
“Wombanizing”
Minutes feels like hours.
Nothing makes sense.
And before you know it, it’s the end of the day.
You’re tired. Your brain hurts.
It feels like you just finished a ton of meetings.
But, at least you shipped 1,000 lines of code, right?
You should be proud.
But, the only thing you feel is hollow.
Did you create? Or did someone else?
That magical feeling from years ago, where did it go?
Rest in peace, flow state.
You were dear to me.
▸ 折叠2条(转推/噪音)
转推2026-09-13 RT @doooyle: If you're mad about the Outbid/Hyrox hype you're missing the lesson
The best business is getting attention for other business…
转推2026-09-13 RT @onlinedopamine: a lot of dev types are suffering from cognitive dissonance
historically, the deal was simple:
> get good at coding
>…
Computers of old were dripping in style. We need to bring all this to Omarchy. I want vintage Commodore 64 and Amstrad 464 themes. Complete with authentic boot sequences. Let's nail the whole heritage pantheon to perfection.
引用@jankeesvwMade an MS-DOS inspired Omarchy theme with my friend @timokuilder. You'd think this much blue gets tiring, but it's honestly really nice to work in. https://t.co/keEy2R5x9V↗
We have big plans for the HEY clients, but I'm also just finding that letting my agent deal with it directly is the most magical interface. Search is amazing, important stuff never slips through the busy box, and all email is easily organizes. Agentic email is wild!
引用@rblalock#OmarchyLinux + Hey App weekend fun. I've loved Superhuman for a long time. HEY is pretty close. The HEY CLI unlocks what I've been missing. So spent some time today building what I want...integrating in to a local agent, etc.
I think the HEY CLI … ↗
Don't get flustered. People are rude, planes are late, and the elevator is broken with no special attention to you. Just smile, rebook, or take the stairs.
Omarchy Quattro has been downloaded more than 300,000 times in the first month of release! Absolutely astounding number. Not sure there's even been a Linux release quite like this before? https://t.co/2Bb7BOA74N
Start with the GDPR and DMA. Flush 'em whole! Then reboot a much, MUCH smaller effort to capture the good notes in a single page with such clarity that not a single lawyer or compliance apparatchik is paid for implementation.
引用@UnamunoAgain🇩🇪 Friedrich Merz, canciller alemán:
“Debemos desregular todos los sectores. Pido un borrón y cuenta nueva regulatoria.
Las pequeñas correcciones legislativas no son suficientes. Necesitamos revisar sistemáticamente toda la legislación vigente de la UE” … ↗
转推2026-09-14 RT @DanWahlin: Putting the finishing touches on the Learn Omarchy app and package. Website is live now but still finalizing some package d…
转推2026-09-14 RT @Tibbzzee: Just set up my Omarchy voice dictation (VoxType). Just hold F9 and talk and it types it out. Runs on your local machine. No c…
Marc Lou@marclouindie395.1K粉 · 9条推进HYROX身体广告的线下履约，并继续记录这场注意力实验的模仿与扩散。
引用@vanshuETHI'm selling my body.
On October 07th, I'm attending my first CRYPTO event in Singapore
You can advertise your startup/brand on 13 spots of my dress:
- Your website listed on https://t.co/JwMn4JF0C8
- Your Logo printed on my dress on event day
- I'll … ↗
引用@marclouI just wrote my first book 📚
I talk about failing for 7 years, depression, and negative thoughts. It's not really sexy, but I want to share the things that don't go viral on social media.
It's free. Please let me know if you enjoy it, that means a lot to me … ↗
I watch his mini VLOG everyday now.
They pop up first in my feed. I see extreme kindness in him. Somehow it reminds me of my wife.
He makes me realize I need to be more grateful.
His brain cancer is a constant sword of Damocles. Yet he’s positive, build side projects, spend time with friends, and make a daily videos.
Maybe I shouldn’t stress out for that workout I skipped?
Maybe I shouldn’t feel guilty if I ate a little bit sugar?
Maybe I shouldn’t feel bad not working today?
Keep up the daily videos please @SudoKidTV. They’re my daily reminder to live life to the fullest ❤️
My grandpa always pushed me to exercise.
He was very athletic when he was young. When we meet, he pats my muscles and say “Look at that! You’d be a fine soldier!”.
I called him to tell him I sold my body for a generous amount of money.
He was baffled. He asked me to explain multiple times until he understood.
Then he said we are lucky to have this internet thing. I couldn’t agree more.
Before hanging up, he asked me to spell it.
“H. Y. R. O. X.”.
He took note and he said he’ll watch YouTube videos to discover this new sport.
The race is in 6 days. I’ll make you proud, Papi ❤️
Paphos is a little place but packed with Twitter people
In the last 2 weeks I’ve been recognized 6 times I think.
Many people here are either indie hackers, remote workers, or ambitious entrepreneurs. Even though it’s a little town, everyone is working on cool stuff.
Nice meeting you Darshan, and safe flight back to Berlin!
引用@WeirdoWizardLook who I randomly stumbled into at Paphos airport.
I’ve been taking inspiration from @marclou's work for a few years now. Never expected to run into him here.
What a perfect way to end my Cyprus trip. https://t.co/QqNaZmaF2W↗
转推2026-09-14 RT @marclou: My grandpa always pushed me to exercise.
He was very athletic when he was young. When we meet, he pats my muscles and say “L…
转推2026-09-12 RT @marclou: I invested in a mini-home gym.
- Bike for zone 2 & 5 cardio
- 2x 32 kg kettlebells for strength
It cost ~$2,000 and covers a…
转推2026-09-12 RT @marclou: Around the time I launched this SaaS, OpenAI and Anthropic started indexing the website.
ChatGPT and Claude are now crawling…
Adam Lyttle@adamlyttleappsindie58.5K粉 · 9条用加载页跳出、难度坡度和会话时长继续迭代网页游戏。
I’m not selling anything on here. Just sharing what I’m working on.
And I think that’s the best way to do it.
I actually have no real business model with my current web game experiments. It’s fun, I enjoy it, and I think it has potential to lead somewhere.
There is a lot of emphasis on monetization and productisation with immediate returns.
But sometimes you gotta just do something you enjoy. Put the reps in. Explore and have fun.
I’m a big believer that you follow that thread and something good comes of it.
Take app development for example:
I started building apps as a creative outlet with new intention of building a real business or doing it full time.
Actually that was never a dream I had because I didn’t even know it was possible.
This has all been a very pleasant byproduct of following my passion and creativity
If you’re building web games don’t underestimate the power of the loading screen
22% bounce rate vs 38%
Factors which contribute to a higher bounce rate:
1. Visual appeal
Showing the game or a scene from the game while loading seems to be best
2. Loading time
Work hard on optimising the size of the initial load. A longer load time = more people bouncing before they’ve even played it
3. Smooth loading
The progress bar should track every byte downloaded not every item. Otherwise you’ll get long pauses in your progress bar that feel like nothing is happening. Adding a spinner to the loading bar also helps to show something is still happening
Every person who quits on your loading screen is a person less who plays. It’s the highest drop off point by far.
Spending a little extra time and attention here is important.
You may also wonder why every game has a PLAY button that shows once the loading is complete (or before the loading)
On mobile web browsers sounds won’t start until the user has initiated an interaction on the site. So if you don’t have a play button before the game starts… no sounds will load until the player starts playing the game. Not ideal.
And finally: when it comes to building the loading screen. Ask your agent to create a static loading screen that isn’t part of your game files or js. We want the loading screen to load before even the js files are downloaded
Bumped up the difficulty ramp and have seen an immediate increase in game session times
I think the balance is not making it too easy. But not making it too hard.
There’s a middle ground to be found. Having a leaderboard lets you see the average times played which helps with targets.
It sounds like the concern is a decentralised ai botnet???
Reading between the lines I suspect the only way these ai swarms were stopped so far was because they’re centralised
Anthropic and OpenAI can just stop them by shutting down the service or filtering requests when they’re detected
But I guess the (legitimate?) concern is that open weights can self propagate… so the same type of behaviour we’ve seen can’t be stopped as easily????
Wild
引用@rynorhnholy shit. dario says that within 6–12 months an agent swarm with the same kind of misalignment we’ve already seen could take over a large portion of the internet, maintain a persistent botnet and do hundreds of billions in damage.
and apparently less severe … ↗
There is a big drop off of players by level 20. I think that’s because the game is kinda repetitive by that stage
You get turbo at level 2
Drones at level 10
Then no new variety in gameplay. Just more challenging levels with tighter timeframes and longer delivery chains
I’m not too sure what other game dynamics can be added. Still thinking about this.
In 24 hours Delivery Dan has become my most played game
Beating games that have been available for weeks
Assumptions I’m taking from this:
You don’t need photorealism to make a game interesting
Lighthearted fun games are more accessible
Games with a protagonist are more compelling
Blending video + gameplay works (I’ve had no complaints about it not working correctly)
All this talk of GPT-6 being nerfed I wanted to test for myself:
Some prompt
Sep 5:
26 mins
10 million tokens
Sep 11:
30mins
8 million tokens https://t.co/dzQLV1fq0l
引用@SahilExecGPT-6 Astra going from a 16-second video to just 4 seconds with the same prompt is a pretty noticeable difference.
If the results really changed this much in just a few days, I can understand why people are calling it nerfed. 😭
The interesting part is … ↗
Tony Dinh@tdinh_meindie201.7K粉 · 1条从创业低谷复盘延伸到高能动性、环境改变与主动跳出舒适区。
I strongly believe two things:
1. The human brain is incredible and is capable of doing amazing things. You just need to have the right mindset and be high-agency
2. The easiest way to change your mindset is to change your environment (where you live, work, routines) and get out of your comfort zone
引用@marclouGet a job.
Entrepreneurship is a mental game. You need extreme resilience and optimism.
I know it because I spent 5 years in the wrong state.
Entitled to my hard work.
Envious of others' success.
Emotional when things don't go as planned.
So I quit in … ↗
Jon Yongfook@yongfookindie171.8K粉 · 2条观察身体赞助可能产品化，也继续用冷幽默判断SaaS周期。
I don’t know what @marclou started but I think at the end of this there will be a billion dollar company that does on-body sponsorships for any kind of activity.
引用@vanshuETHI'm selling my body.
On October 07th, I'm attending my first CRYPTO event in Singapore
You can advertise your startup/brand on 13 spots of my dress:
- Your website listed on https://t.co/JwMn4JF0C8
- Your Logo printed on my dress on event day
- I'll … ↗
Two thoughts watching this:
1. Insane timing. This is going to blow up.
2. This is how every launch video should be done.
引用@contentrewardsWhat if your marketing strategy was just “press this”?
Content Rewards V2 is live. https://t.co/nTu1LF5fDH↗
▸ 折叠1条(转推/噪音)
转推2026-09-12 RT @higgsfield_ai: This is the best motion design video published on X.
Completely built with GPT-6 Astra + Higgsfield Plugin that took ov…
"Outrank backlinks burns your domain"
I saw a few people sharing that last week
(always with an agenda btw, mostly competitors)
i looked at every site with Search Console connected
grouped by backlinks received
here is median Google clicks change since joining
0-9 links: +24%
10-49: +14%
50-149: +42%
150-399: +62%
400+: +115%
my read: if your site receives more backlinks from Outrank ...
👉 it grows more
some connected sites have SEO decreasing traffic (orange below), but most were going down before they started using Outrank. AND again, the more backlinks received by Outrank, the slower the decrease is
Outrank is a net positive for your domain
I run it myself on ALL of the websites in my portfolio
tesla are good but nobody told I would need to stop every 30 minutes to review changes from Claude and start new work 😂 https://t.co/lxiUyltXCs
把开车途中频繁停下复核Claude改动当成Agent工作碎片化的笑话。
▸ 折叠3条(转推/噪音)
转推2026-09-14 RT @virgilerietsch: my SEO since using Outrank :
search console clicks did +25% in the last month btw
you should check them out
https://…
转推2026-09-13 RT @wickedguro: Not so long ago, I got the @tibo_maker newsletter, and it was a really interesting insight.
He said to create a new domain…
转推2026-09-12 RT @pbteja1998: I feel like it's inevitable for https://t.co/WruSSIWaS8 to get so big...
The product is great and the market is huge and…
I’m really trying not to work too much on the weekends and spend more time with the family.
Yet that founder brain of mine won’t shut up. Always prompting myself internally to deal with situations, real or imagined.
It’s quite a challenge to work on this. 🙃
This post of a dev enjoying being a dev is causing a lot of people big feelings.
We software people are quick to identify with stuff, aren’t we…
引用@mehulmptwrote 2 big code files by hand today, single-use scripts.
man these LLMs are total slop machines, even models like gpt 6 astra.
i almost forgot how beautiful and succinct code can be ↗
Marc Köhlbrugge@marckohlbruggeindie88.2K粉 · 4条思考垂直harness为何仍给创业公司机会，并讨论Tesla API与护城河。
Here’s why I’m not convinced the big AI labs will capture the whole software market any time soon
引用@marckohlbruggeI don’t think it’s about the models. It’s about how they interact with the world and how the world interacts with them. In other the words, the harness
An open-source model in the right harness outperforms a frontier model in a mediocre one
The question is … ↗
Why is nobody building on the Tesla Fleet API?
I feel like this might become a huge platform as the cars and FSD becomes more powerful
https://t.co/HOnRB2ecGM
Thank you so much Virgile 😭 that's super nice of you!!
Only 1.5 month remaining to secure our last sponsor spot 👀
It's gonna happen with or without you 😎
引用@virgilerietschcan't believe there is still a sponsor spot left for 2nd edition of Uneed Residency !
drives me crazy because Thomas is building a great community and is the kindest human being
(he can also drive millions of visitors, you're missing out by not sponsoring … ↗
Most people have no idea how feature full Uneed is
🎲 Daily lottery
💬 Private inbox
🤖 Free MCP + API
👥 Product co-owners
❤️ A discord community
🏠 3 residencies per year
✍🏻 Testimonials gathering
👀 450+ alternative pages
🚀 Launch platform, of course
💵 Discount codes management
💌 Newsletter with 26,000 subscribers
🏆 Streaks leaderboard (more to come on this)
📝 Product lists: create, share (or not) lists of products.
📈 Domain rating tracking with email alerts, history, badges
🖼️ Customizable widgets to display testimonials on your website
🐦 FULL social network with feed, integrations, polls, img+videos, product updates etc
These days, my focus is on transforming Uneed into the biggest founders hub and growth platform on the internet.
I want the platform to be alive and useful, even outside of your launch day. I have BIG PLANS for this before the end of the year.
I takes time, and as always people tell me I'm crazy, but so far all the light are green 🥳
Here's my take on Marc's move, because you've all been asking for it 🤪
I wouldn't have done this myself.
I'm not 100% sure why, because as @screenfluent said, this is no different from athletes having brand logos on their clothes. I watch cycling all the time, and they're literally human ads lol
I think it's mostly cultural: in France (maybe in the EU?) we hate advertising. I think most of us don't even know how it works exactly 🤷🏻♂️
But after thinking about it for a few days, I can't help but think:
- There's nothing "bad" about it. No one is harmed.
- It's a genius move that will probably make Marc break out of our tiny X bubble
- It's funny
- Generally speaking, I think Marc is once again pushing the ceiling of "indie hacking" for us. @jonathan_wilke literally showed us a whole new world of possibilities, and Marc is taking it to the next level
I can't wait to see how far it will go 👀
引用@marclouI'm selling my body.
On September 19th, I'm competing in my first HYROX race in Turkey.
You can advertise your startup on 10 of my muscles:
- Your website is listed on https://t.co/49RnsFBVzP
- Your logo will be tattooed on my body on race day
- I'll share … ↗
Simon Høiberg@SimonHoibergindie162.8K粉 · 5条集中担忧前沿放缓走向监管俘获和开放权重受限。
New 2028 prediction:
You work for the AI Cartel - or you're not in business.
引用@DarioAmodeiWe Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, … ↗
My big fear is that current open-weight models will be "banned" or become illegal to use altogether. Which would outright force us to use only the handful of AI providers the government can control.
It's probably unlikely but definitely not unimaginable.
This would be terrible. I'd either have to accept conducting illegal business activities or move to Hong Kong or Shenzhen and continue there (assuming I can still use the Chinese models there).
引用@SimonHoibergThis is regulatory capture. They're trying to prevent China from keeping up by distilling frontier models and preventing smaller startups from using AI not controlled by the government.
And OpenAI & Anthropic are happily pausing to wait for these regulations … ↗
This is regulatory capture. They're trying to prevent China from keeping up by distilling frontier models and preventing smaller startups from using AI not controlled by the government.
And OpenAI & Anthropic are happily pausing to wait for these regulations to be laid out so they can keep monopoly.
This is a huge threat to open-weight sovereign intelligence.
Either this, or something terrible truly did happen that we're not aware of, but I think it's the former.
引用@etnshowBREAKING: 40 British MPs have signed a letter to Prime Minister Andy Burnham calling for a ban on the development of superintelligent AI. https://t.co/zH04Xdw1qq↗
i think people will read this and think "become an influencer" when really it's "become an expert".
nobody trusts influencers. but they trust people who've put in outsized efforts to becoming disproportionately knowledgeable/performant on a topic.
(A+ reply from @kevinrose) https://t.co/eg0uBONYJ9
the initial commit /seo skill also does periodic lighthouse checks and fixes the issues for you.
you never ask it to do those. it just knows when it's time and when it's the highest priority based on all the things it *could* do in a given turn.
https://t.co/tqbgttXTJohttps://t.co/t8lgPdtsqV
引用@einarvollsetSomeone should build a one button microwave. It only does 30 seconds at full power. No other button is ever needed on a microwave. Thank you for your attention to this matter. ↗
no downtime if you've got 6 projects, each with 5 worktrees running in parallel across 4 claude 20x max subs. 🤓
(unrelated: i need a nap)
引用@SebastianRoehlIt's super hard to focus while coding with AI. Nobody talks about this enough. There is SO much downtime, especially with a slow model like Astra. You send a prompt, and now your brain screams at you: do the next prompt, switch to the other project, tab to … ↗
用多项目、多worktree和多订阅调侃并行Agent对注意力的反噬。
▸ 折叠1条(转推/噪音)
转推2026-09-13 RT @james406: STOP ASKING YOUR KIDS "HOW WAS SCHOOL."
they will say "fine" every single time.
a child psychologist told me to replace it…
Andrea Bosoni@theandrebosoindie64.0K粉 · 1条追问B2B SaaS用TikTok或Instagram获客是否存在可复现成功案例。
I’d be really curious to know if there are any B2B SaaS companies that used TikTok or Instagram to get customers. Recently I met a founder who’d been convinced by an agency to spend $$$ on a UGC program on there and the results have been close to zero. Which seems obvious to me because they’re consumer focused platforms. But maybe I’m missing something and there’s someone killing it out there.
Multiple income streams is a good idea.
Multiple apps is not.
I have:
> SaaS
> Boilerplates
> Courses
> X creator
> Books
> Music royalties
> Sponsored content
> Property
> Stocks