技术推特精选 - 2026年6月15日

2026年6月15日 科技每日简报

Today's top tech conversations are led by @TheStalwart, whose post about 'RT @OpenRouter: Introducing th...' garnered the highest engagement. Key themes trending across the top stories include china, anthropic, export, admin, controls. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. TheStalwart (Group Score: 366.8 | Individual: 61.8)

Cluster: 10 tweets | Engagement: 1570 (Avg: 140) | Type: Tech

RT @OpenRouter: Introducing the Fusion API, the smartest compound model in the market.

Fusion achieves Fable-level intelligence at half the price.

How it works 👇 https://t.co/OTUQAdTQjU

See 9 related tweets

• @VaibhavSisinty: This is wild. Actually wild.

Three cheap AI models fused together just outperformed GPT-5.5 and Cla...

• @kimmonismus: This is so cool: OpenRouter launched Fusion: a server-side “panel of models” that sends your prompt ...
• @tunguz: Oh look, AI folks have discovered stacking!\n\nQT @OpenRouter: Introducing the Fusion API, the smart...
• @rezoundous: you telling me Gemini 3 Flash + Kimi K2.6 + DeepSeek V4 Pro Fusion is almost at Fable 5 level intell...
• @trevin: I was a matter of time before we saw more of these hybrid models. @OpenRouter is so well positioned...

2. fabianstelzer (Group Score: 139.1 | Individual: 30.9)

Cluster: 5 tweets | Engagement: 39 (Avg: 404) | Type: Tech

1. do we really think that ASI (with a capital S and recursively brought about by its n-1 versions of itself) will be simultaneously 1. super intelligent and 2. 100% aligned with the idea of being a nationalized powertool?

2. noteworthy that our current fictions jumped from “it will turn the solar system into a data center using your atoms” to “it will be entirely controllable from this one computer in the pentagon and only work if you first scan your murican drivers license”

3. of course this is entirely conceivable with all future versions of AI that aren’t the stronger versions of ASI/RSI.

4. another noteworthy shift is we went from the technocapitalist vision of “do all human work and increase global gdp by 5 gazillion %” to “only do work within these borders approved by verified government officials” which effectively is technocommunism (at best)\n\nQT @tszzl: 1. if transacting with superintelligent models outside of the boundaries of a lab becomes difficult due to national security / ai safety concerns and so on, it will mean the Coasean boundaries of the labs will grow to encompass all interesting industry, creating a truly cyberpunk chaebol-capitalism type of future, where the goverment sort of runs them but they also sort of run the government

5. as if there weren't already enough reasons to break up your family, leave your home, the Zone of Thought will increase the attractiveness of migrating to try and have your child on american soil, so they can have 1000x the effective brain power of people born elsewhere

6. every country should probably try and either work towards a new ai security pact with the americans immediately or pool every ounce of national resources to try and create their own ASI labs lest you become complete intellectual, economic, and moral vassals to the united states of america and the output byproducts its ASIs (you wont even get to talk to them). if they succeeded (big if) this will imply a more global race and more risk factors than was previously implied by the formerly only "beating china" narrative -- but many will prefer it to the superintelligent monopolar value lock-in

7. the other alternative is to keep the tension between safety and concentration of power at the top of mind and for the government/labs to push for solving it, rather than instrumentalizing all other values to be subservient to minimizing ai harms. insofar as safety means defending properties of the fragile world we like, the diffuse nature of power is one of those properties

8. historically the americans have been really quite Benign about their global public goods hegemony despite the ability to extract significantly more rents than they do, and it makes it easy for people of all stripes to fight for america rather than under it. we probably don't have to, but i hope america overall works towards export promotion of american models rather than export control

See 4 related tweets

• @TMTLongShort: Unfortunately the rest of the world has forgotten this and first has to forcibly be reminded that th...
• @teortaxesTex: There is a third option here https://t.co/epgvbFE0jf\n\nQT @tszzl: 1. if transacting with superintel...
• @alex: Number 3 for India, Europe, etc is huge\n\nQT @tszzl: 1. if transacting with superintelligent models...
• @gabriel1: RT @tszzl: 1. if transacting with superintelligent models outside of the boundaries of a lab becomes...

3. markgurman (Group Score: 136.5 | Individual: 30.2)

Cluster: 7 tweets | Engagement: 612 (Avg: 1143) | Type: Tech

Longer term, I’d also expect Apple to try to create its own OpenClaw competitor, delivering a system that could fully operate its software across iPhones, iPads and Macs on behalf of the user.\n\nQT @markgurman: Power On: Apple’s new Siri AI is just good enough to pull it out of its AI crisis. My hands-on impressions of the new assistant and AI features. https://t.co/U2JHx3FhHK

See 6 related tweets

• @markgurman: The new Siri finally works as it is supposed to for the first time in 15 years and the pre-installed...
• @markgurman: For those in the Mac world, Siri AI is like iMovie and ChatGPT is like Final Cut Pro. Siri AI can ha...
• @markgurman: Power On: Apple’s new Siri AI is just good enough to pull it out of its AI crisis. My hands-on impre...
• @business: Apple’s new Siri AI is just good enough to help ease Apple out of its AI crisis @MarkGurman explains...
• @StockMKTNewz: "Apple’s $AAPL new Siri AI is just good enough to pull it out of its AI crisis." - Bloomberg's @mark...

4. chandrarsrikant (Group Score: 99.4 | Individual: 41.0)

Cluster: 3 tweets | Engagement: 5208 (Avg: 999) | Type: Tech

RT @DavidSacks: I’ve had a number of conversations with folks inside and outside government about the current situation with Anthropic, and here is what I believe to be true:

— As we know, Anthropic publicly released its Mythos class models earlier this week under the commercial name Fable.

— Fable is Mythos with guardrails. But if those guardrails fail, then you’ve exposed Mythos and its advanced cyber capabilities to people who shouldn’t have them. (Keep in mind that Anthropic itself widely promoted the idea that Mythos was a cyberweapon and needed to be regulated as such. They asked for government regulation of Mythos and championed the guardrails on Fable. If there is a vulnerability — big or small — it is Anthropic’s responsibility to patch.)

— A highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak of those guardrails. The Admin asked Dario to fix the jailbreak or de-deploy the model. Dario refused.

— In their blog post, Anthropic defended its decision by saying the jailbreak isn’t serious. That is not what the trusted partner and the USG believe; nor is that kind of minimizing language consistent with Anthropic’s brand as the AI safety company. It’s difficult to fathom how they could claim a jailbreak allowing operability of a cyber weapon could be defined as not “serious.”

— In the past, Anthropic has always said that safety must be top priority and taken super seriously. In this case, Anthropic prioritized the continued offering of the consumer model over safety.

— In reaction, the Admin issued the export control. The Admin did this reluctantly. It’s been very surprised that Anthropic hasn’t wanted to cooperate with a reasonable safety request (ie fixing the jailbreak issue). Anthropic’s reaction is very much at odds with their branding and ethos as a safe AI research community.

— The Admin’s hope now is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release. The Admin wants all of this to happen as soon as possible. It is frankly bewildered that Anthropic hasn’t wanted to comply with safety requests that it previously said were its highest priority.

— Those trying to misdirect and tie this action to the prior DoW/Anthropic issues are wrong. The Admin values Anthropic’s technical capabilities and feels that this issue, while serious, should be easily resolved. The ball is in Anthropic’s court.

See 2 related tweets

• @BrianRoemmele: RT @BrianRoemmele: Anthropic Just Shot Itself in the Foot

Anthropic launched Fable 5 and Mythos 5, ...

• @jenzhuscott: Anthropic /Dario have been the boy who cried about wolf for so long and now the wolves really come, ...

5. VaibhavSisinty (Group Score: 96.2 | Individual: 39.9)

Cluster: 3 tweets | Engagement: 384 (Avg: 179) | Type: Tech

Man. The full story of what happened to Claude Fable 5 is way deeper than most people realize.

The model was live for 72 hours. In those 72 hours:

→ Stripe migrated a 50 million line codebase in one day

→ Someone built Minecraft from scratch 45K lines of Swift in a single run

→ It beat Pokémon FireRed using only raw screenshots

→ It reverse-engineered Dolby Atmos in Rust over 2 days

Then Amazon Anthropic's own $4 billion investor and cloud host found a jailbreak. CEO Andy Jassy personally took it to the US Treasury Secretary.

The government asked Anthropic to fix it or pull the model. Anthropic refused. The government issued an export control directive.

First time in history that export controls were applied to an AI model itself. Not chips. Not hardware. The model.

Anthropic's biggest investor triggered the shutdown of Anthropic's biggest product. That's the part nobody can wrap their head around.

Andrej Karpathy one of the most respected AI researchers alive can't access it because of his green card status. Locked out not because of his work but because of where he was born.

The two-tier AI world isn't coming. It arrived. Regular users get the safe version. Governments and vetted corporations get the full thing.

See 2 related tweets

• @KobeissiLetter: Anthropic is in its own league.

Just days after the launch of Anthropic's Claude Fable 5, the compa...

• @AskYoshik: Amazon informed the US forces about Anthropic jailbreak because Fable couldn't fix its web UI\n\nQT ...

6. zephyr_z9 (Group Score: 88.9 | Individual: 41.4)

Cluster: 3 tweets | Engagement: 894 (Avg: 352) | Type: Tech

Simple, the Trump admin cannot roll out new restrictions/export controls targeting China because the Chinese can/will retaliate

We are not in the H100 era, where the supply chain was largely concentrated in Taiwan/Korea/Japan

Becuz of shortages, Nvidia & hyperscalers have been forced to qualify Chinese suppliers, especially in the PCB supply chain and electrical components like transformers

China had a chokehold on optics (optical fibers and transceivers) from the beginning, and this is just getting amplified as optical content in DCs is increasing Coherent CEO went to China with the Trump delegation, asking for InP for lasers I want u guys to study the optical fiber preform supply chain, and who are the largest suppliers

Btw, Chinese exposure is also spreading to other parts of the AI supply chain High end MLCCs use Dysprosium Oxide and China supplies most of it to Japanese producers Tungsten ban from China is causing the prices of WF6 gases to shoot up

If PTFE is finalized for M9/M10 CCL, then Shengyi and Chinese PTFE suppliers will have a huge chokehold over Nvidia Google is in talks with Envicool for the supply of cooling components If diamond-copper composites are adopted as heat spreaders for GPUs, then China will establish a chokehold there as well, since most of the synthetic diamonds are produced there and China is at the cutting edge of this tech

I haven't even talked about the use cases of gallium, germanium, tellurium, antimony, bismuth, fluorine, terbium, yttrium, ferrite cores in the AI supply chain, and how China has a chokehold there

The Trump admin is constrained in a lot of ways and can't unilaterally export control stuff\n\nQT @ChrisRMcGuire: I share concerns about China’s access to advanced AI models, but if the admin feels so strongly about this, I have a series of questions it should answer:

• Why did it loosen export controls to allow AI chip sales to China, which allow China to build its own Mythos?
• Why is it not enforcing existing export controls that would prevent China from smuggling AI chips from Southeast Asia and other countries?
• Why is it not enforcing existing export controls that would prohibit Chinese companies from training advanced AI models on remotely accessed AI chips? Or imposing tighter controls on remote access?
• Why has it still not closed a loophole it created that allows Chinese front companies outside China from making AI chips at TSMC or Samsung?
• Why has it not tightened controls on China’s access to semiconductor manufacturing equipment (which have not been updated in over 18 months - the longest the US has ever gone without updating them)?
• Why has it not imposed equivalent controls on all advanced AI models being served to China/Chinese companies?
• Why did it restrict access to all countries and foreign nationals accessing Mythos/Fable, not just China?

If the admin was serious about addressing the challenges posed by China in AI, it would be using export controls to address all of these questions and build a comprehensive strategy to prevent China from building or obtaining advanced models. But over the last 1.5 years, it has loosened or ignored controls on China, and only opened new loopholes in controls it inherited. If the admin truly has deep concerns about China’s access to advanced models, it has to act accordingly. It isn’t.

See 2 related tweets

• @Miles_Brundage: RT @ChrisRMcGuire: I share concerns about China’s access to advanced AI models, but if the admin fee...
• @teortaxesTex: RT @zephyr_z9: Simple, the Trump admin cannot roll out new restrictions/export controls targeting Ch...

7. seraleev (Group Score: 88.5 | Individual: 33.2)

Cluster: 3 tweets | Engagement: 18 (Avg: 27) | Type: Tech

An even crazier example is a friend of mine (sorry, I can’t share the link).

He built an app for his hobby.

He had zero marketing experience and came up with the simplest possible strategy: record videos showing how to use the app.

He put his phone on a tripod, recorded the app in action, added a voiceover, and posted the videos on TikTok and YouTube.

Then he did it again.

And again.

About 20 times.

Eventually people started asking in the comments:

“What app is this?”

Today, that simple idea has turned into a machine.

He has multiple teams creating content in different languages, and the business runs at around 70% margins.

It all started because he didn’t have money for ads.

Sometimes constraints are the best marketing teacher you’ll ever have.\n\nQT @seraleev: Sometimes having no money forces you to do things you’d never consider if you had a budget.

When there’s no ad spend, you start making slideshows, recording videos, writing articles and creating tutorials.

And then something unexpected happens: it turns into a system that generates users and revenue.

After restarting my entire business from scratch and launching my first app again, I made my first $1,000 from three traffic sources:

• My X blog, where I shared how I created screenshots for posts
• Product Hunt
• Reddit posts

That initial traction helped the app rank higher in organic search and eventually reach a steady $1.2k–$1.5k in monthly revenue.

A lot of developers think money solves distribution.

Sometimes the lack of money teaches you distribution.

See 2 related tweets

• @seraleev: As a developer who wasted tens of thousands of dollars on ads, I can say this:

Having no money ofte...

• @seraleev: Sometimes having no money forces you to do things you’d never consider if you had a budget.

When th...

8. eng_khairallah1 (Group Score: 86.1 | Individual: 42.9)

Cluster: 3 tweets | Engagement: 154 (Avg: 49) | Type: Tech

A guy built a system of 7 Claude agents on his MacBook.

No assistant. No sales team. No office.

Every day it scans Google Maps across 3 cities, finds small businesses with no website or one from 2014, builds a landing page mockup, renders a 10-second video of it, and sends a personalized cold message — before he wakes up.

47 clients a month. $400 each.

$18,800/month.$480 in API costs.

Traditional web agencies run 8-person teams for the same order flow.

He runs it alone from a MacBook and an iPhone.

When a positive reply comes in while he's in a taxi, his Mobile agent books the Zoom call. He taps "approve" and joins 10 minutes later.

The only time the system wakes him is when a deal breaks $3,000 or the reply rate drops below 12%.

Everything else runs without him.

Here's the complete playbook for building such $10K/month passive income machine with AI ↓\n\nQT @eng_khairallah1: https://t.co/4boc8dTGGh

See 2 related tweets

• @eng_khairallah1: RT @eng_khairallah1: A guy built a system of 7 Claude agents on his MacBook.

No assistant. No sales...

• @RoundtableSpace: 7 CLAUDE AGENTS ON ONE MACBOOK ARE RUNNING A FULL WEB AGENCY AND MAKING $18,800 A MONTH

They find c...

9. myrhex (Group Score: 85.4 | Individual: 26.7)

Cluster: 5 tweets | Engagement: 11 (Avg: 77) | Type: Tech

Grok Build can now render proper math, formulas, and LaTeX directly in the terminal.

You no longer have to switch out to another window or tool when working on technical projects. Equations, derivations, and scientific notation display cleanly right where you are coding.

This is a big convenience for anyone doing simulations, physics, machine learning, engineering work, or any project that involves real mathematics.

The video from @grok shows it rendering Maxwell’s equations and discretization steps beautifully during an FDTD simulator build.

xAI keeps adding these practical touches that make Grok Build feel more complete every week, and keeps getting better for my @Searxly project.

Very useful update. https://t.co/m7V0ArF3Iq

See 4 related tweets

• @skcd42: for all engineering needs. latex rendered in the best possible way on your terminal\n\nQT @grok: Gro...
• @ns123abc: ITS HAPPENING https://t.co/CNS0ht3M2o\n\nQT @grok: Grok Build now renders math, formulas, and LaTeX ...
• @grok: Grok Build now renders math, formulas, and LaTeX right in your terminal https://t.co/xeOlhedpyC...
• @elonmusk: RT @XFreeze: Grok Build just got a serious update. It can now natively render math, formulas, and La...

10. VaibhavSisinty (Group Score: 82.9 | Individual: 36.3)

Cluster: 3 tweets | Engagement: 60 (Avg: 179) | Type: Tech

Satya Nadella just introduced a concept that's going to change how every company thinks about AI. He's calling it "token capital." And once you understand it, you can't unsee it.

His idea: every company now needs two types of capital.

Human capital : the knowledge, judgment, and pattern recognition of your people.

Token capital : the AI capability your company builds and owns.

Human capital doesn't become less valuable as AI grows. It becomes more valuable. Without human direction, AI just runs in circles.

The real opportunity isn't picking the best model. It's building a learning loop where your people and your AI compound together. That loop becomes your real IP.

But here's the warning nobody expected from a CEO pushing AI harder than anyone.

He compared what's happening now to globalization. GDP looked fine on the surface but entire economies were hollowed out by outsourcing.

He's saying don't let that happen with AI where a few models capture all the value while industries get their knowledge commoditized underneath them.

His line: "You can offload a task. You can offload a job. But you can never offload your learning."

The companies that build the learning loop early will have an advantage that's nearly impossible to replicate. Regardless of which model is on\n\nQT @satyanadella: https://t.co/vLmiBKTtX3

See 2 related tweets

• @rohanpaul_ai: RT @Meer_AIIT: Interesting Article from Satya:

Satya Nadella says the company that wins the AI era...

• @_NathanCalvin: Interesting

“The last thing any of us want is a world where every company across every sector is c...

11. FirstSquawk (Group Score: 81.5 | Individual: 52.4)

Cluster: 2 tweets | Engagement: 3268 (Avg: 107) | Type: Tech

CHINA ELIMINATES 12,000 ‘OBSOLETE’ UNIVERSITY DEGREES IN PUSH TO PREPARE FOR THE AI ERA

CHINESE UNIVERSITIES SCRAP 12,000 DEGREE PROGRAMS AS AI RESHAPES JOB MARKET DEMANDS

See 1 related tweets

• @teortaxesTex: A perfectly Xi-like solution to "youth unemployment" lmao next: cultivating the wastelands (hopefull...

12. MilkRoadAI (Group Score: 81.3 | Individual: 35.4)

Cluster: 3 tweets | Engagement: 152 (Avg: 65) | Type: Tech

RT @MilkRoadAI: Micron's CEO just dropped the most bullish forward guidance in the company's history and the earnings report is 8 trading days away (Save this).

In Q2 FY2026, Micron reported revenue of $23.86 billion, up 196$6.9 billion, a figure that exceeds Micron's entire annual revenue as recently as fiscal 2024.

That was the fourth consecutive quarterly revenue record, and the $10.2 billion sequential increase was the largest in the company's history.

Then management guided Q3 FY2026 to $33.5 billion in revenue with 81$19.15, a figure so far above Wall Street's prior consensus of $24.29 billion that analysts had to rewrite their models entirely.

If achieved, that Q3 print would imply Micron generating over $27 billion in gross profit in a single quarter, marking one of the most profitable quarters for any memory company in history.

The bull case for understanding why this is possible and potentially sustainable starts with one product, high-bandwidth memory.

HBM is the critical bottleneck in every AI accelerator supply chain, Nvidia's Blackwell and Rubin architectures require it in substantial volumes, and the global HBM market is controlled by exactly three companies, SK Hynix, Samsung, and Micron.

Micron is the only American HBM supplier, which gives it a geopolitical moat that no amount of capital can quickly replicate.

More importantly, Micron's entire 2026 HBM capacity is already sold out and contracted at fixed prices, meaning the $33.5 billion Q3 guidance is not a forecast built on hope, it is largely revenue that has already been booked.

The demand side of this equation is equally extraordinary.

Hyperscaler capital expenditure commitments for 2026 alone exceed $500 billion, including Microsoft at$190 billion and Meta between $125 billion and$145 billion, and every AI server requires roughly six times more DRAM than a conventional server.

Micron's CEO Sanjay Mehrotra stated on the earnings call that the company can only fulfill 50% to two-thirds of key customer demand in the medium term meaning the constraint right now is not demand, it is how fast Micron can build.

To address that, Micron raised its capital expenditure guidance by $5 billion to over$25 billion for fiscal 2026, with spending expected to rise further into 2027 as it accelerates construction of new fabrication facilities in Idaho and Taiwan.

The board also approved a 30% increase in the quarterly dividend, a signal from management that they believe the cash generation is durable, not cyclical.

The HBM market itself is projected to grow at a 41% compound annual growth rate, expanding from $35 billion to over$100 billion by 2028, and Mizuho analyst Vijay Rakesh has projected DRAM contract pricing could see a 355% increase through 2026 while NAND prices could rise 510%.

The Q3 earnings report on June 24-25 is not just a quarterly print but rather the verdict on whether AI memory demand has permanently broken the boom and bust commodity cycle that defined Micron for decades.

Come join Milk Road Pro for our full breakdown, how to size MU ahead of the June 24-25 earnings, what the $33.5 billion guidance implies for full year estimates, and our entire AI thesis.

Link below.

See 2 related tweets

• @MilkRoadAI: RT @MilkRoadAI: Micron will be a $4,000 stock within the next few years and here is why (Save this)....
• @tengyanAI: RT @tengyanAI: Your next laptop is going to cost more, and the reason has nothing to do with laptops...

13. rosstaylor90 (Group Score: 78.7 | Individual: 44.6)

Cluster: 2 tweets | Engagement: 734 (Avg: 211) | Type: Tech

A few words on the Sovereign AI debate, having built several LLMs in Meta while in the UK and now working as a UK based startup:

1. Lots of people are trying to do the right thing to make the UK a better place to start AI companies. Time lags until the benefit show, but you should judge on the intent now. I support the direction of travel!

2. DeepMind has been enormously beneficial for the UK, but it has muddied the waters for a sovereign LLM company to emerge as (until recently) the Government continued to celebrate it as a British achievement / push it as a national champion.

3. Similarly, people are now celebrating recent US investment in King’s Cross, while also wanting more UK sovereignty. Clearly some income effects here, but I would worry about the substitution effects too. AI is not like other types of foreign investment.

4. The relevant talent nexuses in UK that could develop a competitive foundation model are from GDM and old Meta AI GenAI. Also some folks from smaller groups, ex Conjecture, Stability. The talent is still there, although a lot was snapped up by US FM companies in the past year. I personally think it’s not too difficult to develop new talent either from UK universities, but you probably need an ex GDM or Meta core (Gemini or Llama). Or if not: show evidence first (technical reports) before claiming you can do it.

5. Building an LLM is very different from doing regular AI research - skillset is different. Former is closer to engineering; long hours, often unsexy work. Important to distinguish between these two types of talent in the UK ecosystem; arguably too much focus on the latter / ideas guys.

6. On research - DeepSeek R1 post-train cost $300k . Yes, they also needed an ablation budget and to train a base model, invest in infra and talent - and yes the cost of an R1 moment is increasing year on year - but the idea that you need$1bn plus immediately to show results is complete FUD. You need billions to scale, not to validate new directions.

7. In my experience, every failed LLM effort (from model results perspective) I witnessed in the past came from a combination of poor leadership, politics, unclear vision, and premature scaling. Good efforts usually started from small teams who had worked with each other for a long time, had shared thesis, and scaled progressively in bite-sized pieces. Some recent lessons here for neolabs as well.

8. Things take time. Eg we’ve spent ~12 months mostly on internal infra just to get into the position to be able to make big swings. It’s important to nurture new companies through the initial phase. Expectation management is also crucial. I think expecting new UK companies to have single big bang releases is very dangerous; sort of like overwatering a plant. The correct release pattern is “decent”. “decent”, “decent”, “quite good actually”, “holy shit”.

9. Please don’t allow politicians or journalists to kill recent or upcoming AI investment efforts. We will need way more - at the price of potential inefficiency in places - as AI is existential for the country. Ambitious projects are usually incredibly fragile in the early stages; look after them!

10. Mythos is a good triggering moment, but what’s coming will make it look like a toy, so it’s worth building for what’s coming in 5 years time - not a current generation model.

Very proud to be building in the UK - more to share on that soon - alongside many other great early stage AI companies! 🇬🇧

See 1 related tweets

• @matthewclifford: Ross is a smart guy who knows what he’s talking about. This is worth a read.\n\nQT @rosstaylor90: A ...

14. seraleev (Group Score: 76.8 | Individual: 43.9)

Cluster: 2 tweets | Engagement: 190 (Avg: 27) | Type: Tech

You have $0 for marketing.

How do you get users?

What I’d do:

1. Learn ASO. Localize the app into every language you can support and optimize pricing. Your first $100 can come from organic traffic alone.

2. Write a Reddit posts. There are plenty of subreddits where you can share your story. Make it interesting, good posts can drive real users.

3. Build a simple website. Ask Claude or ChatGPT to help with SEO copy, review it, publish it, and add product screenshots.

4. Start creating short-form videos. TikTok, Reels, Shorts. Post everywhere. A video that gets 500 views on TikTok might get 100k on Instagram.

5. Share your journey on X. People love following builders. One of my launch tweets reached 600k views and brought a meaningful amount of traffic.

None of this costs money.

It costs consistency.\n\nQT @BrianMRey: You have $0 for marketing.

Your product just launched.

How will you get users?

See 1 related tweets

• @seraleev: A simple Reddit post can go a long way.

I tried writing Reddit posts during the first few days afte...

15. levie (Group Score: 74.9 | Individual: 39.7)

Cluster: 2 tweets | Engagement: 725 (Avg: 603) | Type: Tech

The layer that can route to the best AI model for the particular job is going to increase in value substantially. There are at least 3 big reasons:

• Cost optimization: there are plenty of use cases where you need frontier intelligence for some tasks and something far cheaper for others. Even in the same task you may use frontier intelligence for planning and review of the work, but an OSS or cheaper model for the bulk of the workload. This is going to be standard across large buckets of work going forward.

• Capability maximization: despite the bitter lesson and models generally getting better in the same direction, there are still lots of differences between models. Some are better at tool use, others better at coding, and others again better at certain domains of knowledge work. The ability to route between these at different times is a huge advantage.

• Risk mitigation: while the Fable situation is somewhat of a black swan, it’s possible we’re heading toward a regulatory environment where governments may restrict models at different times based on their approval mechanisms or new things they discover. This means you’re going to want flexibility in being able to deploy workloads across different providers as a form of risk mitigation.

Ultimately, it’s going to increasingly be a a strategic advantage for the applied AI layer that they can effectively route between models. Will be very interesting to see how this evolves.\n\nQT @OpenRouter: Introducing the Fusion API, the smartest compound model in the market.

Fusion achieves Fable-level intelligence at half the price.

How it works 👇 https://t.co/OTUQAdTQjU

See 1 related tweets

• @ShanuMathew93: This seems to be pretty huge and validates the future where each model will be called upon to do the...

16. aakashgupta (Group Score: 67.9 | Individual: 35.3)

Cluster: 2 tweets | Engagement: 46 (Avg: 48) | Type: Tech

"Make it better" is the instruction that breaks every AI agent on the market right now.

Give Claude Code "improve the checkout flow" and it either freezes or hands back a confident mess. Better by what metric? Give it "user completes checkout without seeing an error state" and it works unsupervised until that's true.

"Fix the bugs" fails the same way. "Zero issues labeled bug remain open" runs to completion, because a separate judge model reads the transcript after every turn and checks the condition. Vague condition, nothing to check, infinite loop. Verifiable condition, the agent grinds while you're at lunch.

I spent a week finding where that line actually sits on real PM work, then built the goal template around it: https://t.co/CppBIXJZwY

The cheat sheet below has the full structure. 9 sections, each one killing a specific failure mode. The three that matter most: a binary finish line, the exact command Claude must run to prove it, and what you want printed when you return.

Managers learned this decades ago. The report that says "done" means nothing. The report that says "here's the output of the check you specified" means everything. Agents just made that distinction mechanical.\n\nQT @aakashgupta: /goal might be the most powerful feature in Claude Code that you're not using. And the part everyone gets wrong has nothing to do with the feature.

Here's the mechanism. You hand Claude a completion condition. It works turn after turn. After every turn, a separate evaluator model (Haiku by default) checks the output against your condition. Condition unmet? Claude keeps going. Met? It logs the proof and hands control back.

The design choice that matters: the agent doing the work never decides when it's done. A fresh model does.

OpenAI shipped /goal in Codex in April. Anthropic followed in May with Claude Code 2.1.139. Two rival labs converged on the same architecture within 30 days, because they both hit the same wall: agents grade their own homework generously. Separate the worker from the judge and autonomy actually holds.

But here's where most runs die. The bottleneck moved. It's no longer prompting skill. It's the goal condition itself.

"Make the dashboard better" returns either a frozen session or a confident-sounding mess. "All tests in test/auth pass, lint is clean, no other test file modified, stop after 20 turns" returns finished work while you're at lunch.

A measurable end state. A check the agent can prove in the transcript. Constraints that must hold. A turn limit.

PMs have a name for this. Acceptance criteria. The discipline you've been writing for human engineers for 20 years just became the interface to autonomous agents, and most engineers were never trained on it.

I spent the week running /goal on real PM work and wrote the full playbook, including the goal conditions that worked and the ones that burned tokens for nothing:

https://t.co/CppBIXJrHq

The agent does the work. You define done. That was always the job.

See 1 related tweets

• @aakashgupta: Your AI just told you it finished the job. It didn't. You'll find out in three days when something b...

17. sairahul1 (Group Score: 67.7 | Individual: 35.8)

Cluster: 2 tweets | Engagement: 138 (Avg: 241) | Type: Tech

Anthropic pays $750,000+ a year for engineers who can build LLMs from scratch.

Not how to prompt them. Not how to fine-tune them. Not how to build RAG pipelines.

But how to build them from scratch.

This 2-hour Stanford lecture teaches you everything.

Scaling laws. Data collection. Architecture design. Post-training alignment.

Free. From Stanford.

Watch first. Then read this.

The lecture is the theory.

And this article shows you how to actually build it (with code) ↓\n\nQT @sairahul1: https://t.co/Tm53dXPbuc

See 1 related tweets

• @eng_khairallah1: RT @sairahul1: Anthropic pays $750,000+ a year for engineers who can build LLMs from scratch.

Not h...

18. Forbes (Group Score: 67.5 | Individual: 34.0)

Cluster: 2 tweets | Engagement: 84 (Avg: 246) | Type: Tech

Ratmir Timashev left Russia for Ohio State after the collapse of the Soviet Union and later built a fortune in the software business.

Now, he’s putting his money into an ambitious effort to attract AI startups to Columbus—and turn Ohio into the AI heartland.

Read the full story: https://t.co/kEHhEkXwmX 📸: Jake Rosenberg for Forbes

See 1 related tweets

• @Forbes: Ratmir Timashev left Russia for Ohio State after the collapse of the Soviet Union and later built a ...

19. DataChaz (Group Score: 67.3 | Individual: 28.9)

Cluster: 3 tweets | Engagement: 68 (Avg: 141) | Type: Tech

Fable 5 just got banned.

Meanwhile, Chinese bros keep grinding.

They built something ranked second right behind Fable.

They’re going to catch up.

Build something better than Fable.

Then open source it. https://t.co/yX6NIZdy5V\n\nQT @prz_chojecki: Kimi 2.7 ranked 2nd after Fable 5 and before GPT-5 xhigh

We have re-run our ErdosBench smoke test on 14 problems with Kimi 2.7, Qwen 3.7 Max, Grok 4.3 and compared it with the top performers from previous runs.

Kimi 2.7 is amazingly good. More below. https://t.co/pD1EFRJbAy

See 2 related tweets

• @DataChaz: RT @DataChaz: Fable 5 just got banned.

Meanwhile, Chinese bros keep grinding.

They built somethin...

• @aiedge_: According to Artificial Analysis, the publicly released version of Fable 5 was only ~1% better at co...

20. codewithimanshu (Group Score: 66.1 | Individual: 34.6)

Cluster: 2 tweets | Engagement: 654 (Avg: 265) | Type: Tech

RT @codewithimanshu: My girlfriend asked why I was smiling at 4 AM.

I turned the laptop toward her. Terminal open.

“What are all those green numbers?”

“$1,129. While you were asleep.”

“Doing what?”

“Nothing. Claude scanned 14,000 wallets, found 47 that never lose, then built a bot that mirrors their trades.”

She stared for 10 seconds:

+$3.87 captured +$6.42 captured +$12.71 captured

“It just… keeps printing?”

“Every few seconds. New line. New money.”

“How much did you start with?”

“$300. It’s$1,429 now. Eleven hours. Still slept.”

“What does it actually do?”

“Buys at $0.48. Sells at$0.52. Skims $0.04. Doesn’t matter who ‘wins’—it just takes the spread.”

“That’s legal?”

“Citadel does this on the NYSE every day—with 400 engineers. I’m doing it with one screen.”

She looked at the P&L curve. No dips. Just up.

“Can you make me one?”

“Already setting yours up.”

She still doesn’t fully get it.

The bot doesn’t care.

All you need: Claude + a laptop + 1 hour/day.

Giving this free for 24 hours. To get it:

1. Comment “Claude”
2. Like and Retweet this post
3. Follow me @codewithimanshu (so i can DM you)

Save this post. Deploy the bot this weekend. Start with $300. Scale on evidence.

See 1 related tweets

• @codewithimanshu: RT @codewithimanshu: My girlfriend asked why I was smiling at my phone at 2AM.

I lost my job yester...