顶尖科技推文 - 2026年7月7日

Today's top tech conversations are led by @DeItaone, whose post about 'MICROSOFT CUTS 4,800 JOBS

Mic...' garnered the highest engagement. Key themes trending across the top stories include compute, https, already, internet, about. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. DeItaone (Group Score: 303.7 | Individual: 25.2)

Cluster: 16 tweets | Engagement: 643 (Avg: 486) | Type: Tech

MICROSOFT CUTS 4,800 JOBS

Microsoft is cutting 4,800 jobs, or about 2.1% of its global workforce, mainly across its Commercial and Xbox divisions.

The company said the restructuring is aimed at improving operations and positioning Xbox for long-term growth.

Microsoft added that the layoffs are not driven by AI replacing employees.\n\nQT @DeItaone: $MSFT - MICROSOFT CUTS 4,800 JOBS ACROSS SALES, XBOX: BUSINESS INSIDER

See 15 related tweets

@rohanpaul_ai: Microsoft is cutting 4,800 jobs as AI infrastructure spending forces a harder tech reset, equal 2.1%...
@WOLF_Financial: BREAKING: MICROSOFT $MSFT IS CUTTING AROUND 4,800 JOBS, ABOUT 2.1% OF ITS GLOBAL WORKFORCE

Almost e...

@negligible_cap: *MICROSOFT TO CUT 6,400 JOBS, HALF IN XBOX GAMING OVERHAUL - $MSFT\n\nQT @negligible_cap: *MICROSOFT...
@StockMKTNewz: MICROSOFT $MSFT IS CUTTING 4,800 JOBS, AND ITS OWN HR CHIEF WENT OUT OF HER WAY TO SAY AI ISN'T WHY ...
@tomwarren: BREAKING: Microsoft is laying off 4,800 employees today. Most of the job losses are in Microsoft’s X...

2. teortaxesTex (Group Score: 164.4 | Individual: 34.7)

Cluster: 5 tweets | Engagement: 45 (Avg: 60) | Type: Tech

Very important whether these assumptions hold. https://t.co/oebBt2CYR8\n\nQT @willdepue: A Stargate for Data

Labs are on a trajectory towards >$100B/year of data spend by 2030. As we begin the trillion-dollar compute project, we need to think about the equivalent civilizational-scale effort for the other core ingredient: data.

At the foundation of the scaling revolution is a simple empirical law: deep neural networks improve smoothly, near magically, as you scale two things in proportion — (1) the size of the model and (2) the amount of data you train on. And despite the scaling laws being brutally diminishing, we’ve successfully bitten the bullet of logarithmic scaling with exponentially larger clusters and datasets, and received incredible new capabilities in return.

But this exponential scaling is bound to hit some limits. Oddly enough, compute has compounded fairly smoothly without limit, with trillions flowing into hypercluster buildout. Instead, we’re starting to hit the limits of an exponential demand for data. Gone are the days of being purely in the compute-limited regime, where we had effectively infinite internet data but never enough GPUs, we’re now entering a data-limited regime.

Luckily, this limitation is coinciding with staggering improvements in AI capabilities. Incredibly, we seem to have a real line of sight towards automating a majority of knowledge work with the methods we have today. RL + pretraining, and the data for each, will be generally sufficient to achieve most economically valuable tasks, given some minimal algorithmic progress and continued compute scaling.

In a data-limited world, economic progress & scientific acceleration will be directly bottlenecked by our coverage in each domain. We need to see data collection as imperative, deserving the same civilizational ambition we’ve given compute.

The internet as a one-time subsidy

It’s underrated how much all progress in AI owes everything to the blessing of the internet, this one-time civilizational subsidy to deep learning, decades of unintentional accumulation of a perfect dataset: every book, blog post, image, video, paper, discussion, etc. all digitized and freely available. Without the internet, we’d likely see comparably minimal progress in AI today, and in fact, if you notice where systems currently underperform, it’s almost always a domain where web coverage is limited and data is private, expensive, non-digitized, or non-existent.

But we’re running out of it. There are only about 300 trillion tokens of useful public human text, and the internet doesn’t produce nearly enough new high-quality data to match what scaling demands — we’re soon to hit the limits of public data for pretraining. And though the advent of RL bought us reprieve — chain-of-thought RL needed a new form of untapped data, gradable math & coding tasks, also available online — we’re quickly running dry of hard tasks for RL as well.

Why do we need so much data anyways? Humans learn comparably in far less time, needing just one textbook where language models might need the equivalent of hundreds to learn a new topic. It’s possible we discover methods that are massively more data efficient — synthetic data, data efficient architectures, other exotic algorithms — but fundamental progress is slow and highly unpredictable, and the recipe we have just works today.

And, while I’m wary of getting too deep here, even arbitrary data efficiency can’t replace data that just doesn’t exist in the first place. There’s a massive amount of missing information on the web: the dark matter of the internet — tacit knowledge, undocumented processes, etc. — most of which was never published and lives only inside organizations, the physical world, or just in people’s heads. I’ll leave it here and say, for reasons far longer than I can fit in this post [1], it’s best to operate on the assumption that our insatiable desire for data will continue as it has for the last decade.

There will be >$100B/year in data spend by 2030

We’re not screwed yet, of course. Only a fraction of useful data in the world is on the public internet, the rest is stored inside private datasets, corporations, personal archives, universities, governments, and otherwise. Labs can and will continue to license these private datasets, or create them from scratch, like Anthropic’s book scanning project. And we’ll increasingly task human experts to manufacture new high-quality data, with a large fraction of hard RL training tasks already being sourced this way.

But collecting this data, unlike before, will be expensive. As the free internet dries up and demand for data rises, we should see labs investing equally in data as compute, likely spending a significant fraction of their compute budgets on data. As we see trillions spent on compute, we should also expect hundreds of billions spent on data (human data & collection budgets), given their equivalent importance. And, notably, data spend is already tracking this way: total data spend across vendors, not counting internal lab efforts, is already roughly $7 billion per year. It’s quite reasonable we’ll see >10x by 2030.

Data is the moat

Data becoming increasingly private will also majorly shift the competitive landscape. While compute is a commodity — everyone buys the same chips and builds the same clusters — data really isn’t. The big reason why frontier models have felt eerily similar to one another, until now, is they were trained on substantially the same internet (pretraining data variability across labs seems pretty low). As labs diverge onto more exclusive, manually collected corpora, I think models will begin to increasingly diverge.

OpenAI pulling ahead in mathematics and Anthropic in cybersecurity isn’t an accident. I really think laser-focused collection of high-quality midtraining tokens, custom RL tasks, environments, with dedicated research effort, has driven much of the visible progress in the last year. James Betker has an excellent blog about “the ‘it’ in a model is the dataset”: model architecture and compute buy you efficiency and order-of-magnitude performance, but ultimately, models, of any architecture, are such incredible approximators of their dataset that the core meat of a model boils down to just that, nothing else. Data is a major moat.

AGI long, ASI short

As I’ve tweeted before, I’m confident that, despite the narrative, the data labeling industry will continue to fuel great businesses and be an excellent AGI long, ASI short. The argument is just: By the time the AGI labs no longer need data, it’s probably over for everything else too [2]. In this frame, the last companies left should be the data companies, as the last speck of economically relevant data is sucked in. And these companies are already among some of the fastest-growing companies in history: Mercor, founded three years ago, is rumored to be doing $2 billion in revenue with something like a few million expert labelers under contract.

While these businesses are very non-stationary, what type of data is needed shifts constantly, I don’t think that diminishes their value. The long-tail of the economy is long, and the value isn’t diminishing as you extend farther into more obscure information: as models get more capable, the value of the marginal dataset goes up, not down. Automating a full job means covering its full distribution of tasks, tools, edge-cases, and long-horizon loops. There’s some O-ring logic to it: a dataset that buys a 1% bump can justify a previously unjustifiable collection cost when it’s the difference between a system that does 99% of a job and one that does all of it [3].

The competitive dynamics of the data industry are still evolving but as demand for data is increasingly niche, ultra high-quality, expert-generated, I think we’ll see real consolidation. Again, contra-narrative, we’ll probably see true competitive differentiation built on brand, quality control of data (which, from personal experience, can vary massively), as well as in network effects from the talent networks themselves over time. We’ve already seen rapidly shifting data type demand work in favor of incumbents, benefiting those with early knowledge of where the market is headed.

The binding constraint

It’s truly remarkable that we seem to have the recipe — pretraining + RL — to absorb most economically valuable work, despite being far from a lot of what we expected from “AGI”. The same way chess engines revealed we never needed general intelligence to solve chess, as we originally thought, we’ll soon realize that software, mathematics, and the vast majority of the economy (including physical, just running ~3 years behind!) are the same. If recursive self-improvement or some other algorithmic breakthrough arrives, that’s wonderful, but we really don’t have to wait for it. The binding constraint between here and an automated economy isn’t that, it’s data coverage: every app, workflow, edge case, process, etc. sitting in private stores or someone’s head.

Ultimately, while we make tremendous strides in more efficient model architectures, and clusters like Stargate equip us with zettaflop-scale compute, we really aren’t making rapid progress collecting the data we lack.

We’ll soon live in a world where we have the methods & compute to accelerate scientific progress or economic growth, but not the data. And we’re already there today: frontier models would surely be as good at accounting/many medical tasks/legal advice as they are at software engineering if we only had the same pretraining & RL coverage as we did for code.

I really want to drill this in: The speed at which we automate the economy is going to be directly rate-limited by our ability to collect data about it.

Worth noting that under this assumption, with data as defensible and directly proportional to economic & scientific progress, data should also be considered a national strategic asset like compute. Imagine what we’d do in a world where we had a Manhattan Project-effort for AI and needed to mobilize data collection as a limiting factor. We should be concerned about China, with greater state capacity and authoritarian economic control, being capable of mobilizing data collection at national scale, potentially compounding their economy and scientific output faster than us down the line.

A Stargate for data

I’m leaving my complete ideas for a future post, as this one is already far too long, so I’d really like to pose the question here. Stargate exists because we organized trillions of dollars, international strategy, gigawatts around compute as a fundamental ingredient. What would equivalent ambition look like for data?

Obviously, scaling data collection, a heterogeneous mass of information across the economy, isn’t going to be as clear as scaling compute, as a homogenous infrastructural effort. A core division will be first, coverage — all uncaptured knowledge sitting across the economy/science/physical world and all that simply isn’t recorded — and, secondly, sheer volume in the domains we already train on: more hard math tasks, more high-quality web text, way more coding data, more legal drafts, etc.

I have a post coming soon which breaks down my proposals. There’s a lot of room for creativity. Quickly, we’ll probably want to start with a deep census of what we have and what we’re missing, predict what the 2030 model will still be bad at and work backward to what we should be collecting today. You can probably license a large amount, leveraging high lab valuations to buy datasets or companies altogether. There’s an adversarial nature to a lot of this collection with firms, so there’s lots of engineering to do this correctly. We should go convince important companies to turn off deletion policies, even if we’re not buying from them yet. Data flywheels in consumer products will be massive. Confidential training, government legislation for grant-funded research, running companies at a loss for their data, etc.

We’re headed towards hundreds of billions in expenditure, national prioritization, and major data limitation on the horizon. We have a great opportunity to think creatively about what a megaproject for data would look like: How do we, deliberately this time, construct the next internet’s worth of data?

Footnotes:

[1]: I’ll probably soon publish my much longer post explaining my position on data efficiency and why the value of this data is still pretty high in most worlds regardless of new algorithms.

[2]: The “AGI freeroll” bet: heads you win, tails ASI flips the world upside down anyways.

[3]: We already see a glint of validation of this point, given the data market is strongly tilting towards ultra-high-quality agentic data, rather than unskilled labeling — niche expert workflows, live environments, and evaluations requiring increasingly obscure talent & knowledge — yet shows increasing, not decreasing, revenues.

See 4 related tweets

@giffmana: It's just slightly too long read, but worth it. I basically nodded all the way through, I think ther...
@0xdevshah: there is also a massive data opportunity in deliberately discovering and engineering entirely new sp...
@samsja19: do not delete your production trace, turn them into fuel for your next post training\n\nQT @willdepu...
@gabriel1: RT @willdepue: A Stargate for Data

Labs are on a trajectory towards >$100B/year of data spend by 20...

3. victormustar (Group Score: 157.6 | Individual: 38.5)

Cluster: 5 tweets | Engagement: 374 (Avg: 63) | Type: Tech

RT @TencentHunyuan: 🚀Hy3 is here.

295B MoE. Best in its size class. Rivals trillion-scale flagships. Reliable and affordable for most agentic usecases. Apache 2.0. Friendly for commercial use. FREE API for 2 weeks → https://t.co/EyURKwTdgi

🤗 https://t.co/twqJpqb2SL 📖 https://t.co/4uEkIU1cW4 https://t.co/LLJTfDFNQu

See 4 related tweets

@guohao_li: congrats the Hy3 release @TencentHunyuan! free api usage for two weeks is amazing. if we can get mor...
@perrymetzger: Those who are not paying attention seem to be fully unaware of the rate at which Chinese companies r...
@vllm_project: Spin it up now! 🚀 https://t.co/5KdGTTeCgn\n\nQT @TencentHunyuan: 🚀Hy3 is here.

295B MoE. Best in it...

@Michaelzsguo: Very smart Hy3 allows us to use the API for free for 2 weeks. https://t.co/asRWRDv1Xe\n\nQT @Tencent...

4. Reuters (Group Score: 156.2 | Individual: 35.0)

Cluster: 5 tweets | Engagement: 184 (Avg: 116) | Type: Tech

South Korean chipmaker SK Hynix will launch a US listing to raise about $28 billion, according to regulatory filings, as it capitalizes on the global artificial intelligence boom with one of the world's largest new share sales https://t.co/RSRAjph6cm https://t.co/vQhMmyryyo

See 4 related tweets

@ReutersBiz: WATCH: South Korean chipmaker SK Hynix will launch a US listing to raise about $28 billion, accordin...
@Reuters: South Korean chipmaker SK Hynix will launch a US listing on July 6 to raise about $28 billion, accor...
@ReutersBiz: South Korean chipmaker SK Hynix will launch a US listing on July 6 to raise about $28 billion, accor...
@DeItaone: SK HYNIX GOES BIG ON AI BOOM

SK Hynix is raising up to $28 billion through a U.S. ADR listing, one ...

5. yuhasbeentaken (Group Score: 132.5 | Individual: 34.6)

Cluster: 4 tweets | Engagement: 57 (Avg: 23) | Type: Tech

if you’re wondering how tencent hy3 compares to other open models...

here are the agentic coding benchmarks:

swe-bench multilingual

hy3...75.8
glm-5.2...83.0
deepseek v4 pro...76.2

swe-bench pro

hy3...57.9
glm-5.2...62.1
deepseek v4 pro...55.4

hy3 is not beating glm-5.2 yet but it is already very competitive with deepseek v4 pro across coding and agentic workflows.\n\nQT @yuhasbeentaken: hy3 is out...

5 things the benchmarks suggest:

1 .hy3 is not another cheap open model 2. it is already competitive with frontier models on real workflows 3. it nearly matches gpt-5.5 and opus 4.8 on browsecomp 4. it beats gpt-5.5 on frontierscience-olympiad 5. it shows big gains over hy3 preview across multiple agentic tasks

some numbers:

browsecomp...hy3 84.2 vs gpt-5.5 84.4 >frontierscience-olympiad...hy3 74.8 vs gpt-5.5 73.8 >mcp atlas...hy3 79.1 vs glm-5.2 82.6 claw eval...hy3 68.5 vs glm-5.2 62.4 terminal bench 2.1...hy3 71.7 vs deepseek v4 pro 64.0

5 things that still need testing:

real coding agent performance
long-context stability
tool-use reliability in production
inference cost at scale
actually beats glm-5.2 outside selected benchmarks or it's just hype?

the takeaway?

open models are getting good enough on real workflows while being much cheaper and easier to deploy.

that is the real threat to frontier pricing power.

See 3 related tweets

@yuhasbeentaken: hy3 is already near the top in agentic search...

browsecomp

hy3...84.2
deepseek v4 pro...83.4...

@0xSero: I ran preview as soon as I got my 4x 6000s and it was phenomenal. A bit rough around the edges but w...
@yuhasbeentaken: hy3 changes the open-source model story...

it is a 295b moe model...

but only 21b parameters are a...

6. teortaxesTex (Group Score: 108.7 | Individual: 47.5)

Cluster: 4 tweets | Engagement: 521 (Avg: 60) | Type: Tech

If these scores are real, Tencent has just become one of the leaders of open source. Hy3 is a 295B 21AB model that at least beats GLM 5.1 in blind tests. Might even warrant architecture transition (it's a basic GQA). Good job. https://t.co/g3uSESesXk\n\nQT @xeophon: Tencent @TencentHunyuan dropped the non-preview version of Hy3 and changed their license from the community one (restrictive + not allowed in SK, UK, EU) to Apache 2.0!!! 🙏 https://t.co/OChvsg5zjy

See 3 related tweets

@testingcatalog: Tencent released Hy3 (non-preview), a 295B-parameter MoE open model with 21B active and 3.8B MTP lay...
@zephyr_z9: Pretty strong model from Tencent\n\nQT @teortaxesTex: If these scores are real, Tencent has just bec...
@chris_j_paxton: Not a big model either right? GLM is 2.5x the size. Driving the cost of intelligence down lower and ...

7. thsottiaux (Group Score: 104.9 | Individual: 30.3)

Cluster: 4 tweets | Engagement: 2010 (Avg: 4898) | Type: Tech

Closing this week. Much of the teams building ChatGPT, Codex and OpenClaw will be there and we'll have a few surprises for folks. Hope to see many of you 👀\n\nQT @OpenAIDevs: Applications to attend OpenAI DevDay 2026 are open.

Join us in San Francisco on September 29 to:

• Tinker with what’s new • Swap build notes with other builders • Go deep in technical sessions • Bring your sharpest questions

Apply by July 10:

https://t.co/mda0bFx2Hp

See 3 related tweets

@reach_vb: PSA: Just 4 days left to apply for DevDay 2026!

Join us for a day with the OpenAI community, meet t...

@OpenAIDevs: Ship your OpenAI DevDay 2026 application by July 10:\n\nQT @OpenAIDevs: Applications to attend OpenA...
@dkundel: Still energized from AI Engineer World's Fair! It was so good to see so many familiar faces in one p...

8. RoundtableSpace (Group Score: 103.3 | Individual: 27.9)

Cluster: 4 tweets | Engagement: 126 (Avg: 127) | Type: Tech

Someone making $50K a month with Fable 5 leaked the one-page cheatsheet Anthropic buried in their own docs.

4 effort levels, 50+ subagents, one orchestrator, the entire prompting system for the most powerful model on earth on a single screen.

The barbell that decides if it pays: Fable plans 10%, cheap agents run 80%, Fable verifies 10%. Tell it the why, keep the prompt short and cap it with a stop rule.

Your old Opus 4.8 prompts actively make it worse. Anthropic hid the fix deep in the docs and people sell $500 courses on it. You get the whole thing free.

See 3 related tweets

@DataScienceDojo: 🚨 Fable 5 stops being free tomorrow. One day left before it's fully metered.

The move going around ...

@rewind02: RT @rewind02: Fable 5 on low effort: $3.76 per task. Opus 4.8 on max:$ 13 per task

Pass rate? Fable...

@alex_prompter: RT @alex_prompter: You can make Fable 5 teach Opus 4.8 how it thinks before it moves to pay-per-use....

9. kimmonismus (Group Score: 102.5 | Individual: 29.5)

Cluster: 4 tweets | Engagement: 2760 (Avg: 619) | Type: Tech

Confirmed by OpenAI Tibo: GPT-5.6 Sol Ultra will be in Codex.

Tomorrow is going to be an insane day. https://t.co/Xzeit07YvQ\n\nQT @thsottiaux: @haider1 Ultra will be in codex.

See 3 related tweets

@VadimStrizheus: 🚨 BREAKING: Tibo from OpenAI confirmed that GPT 5.6 Sol Ultra will be in Codex

this will force Anth...

@mark_k: Good news: GPT-5.6 Sol Ultra is coming to Codex! This is the equivalent to GPT Pro, which many of yo...
@yuhasbeentaken: 3 things we know about gpt-5.6...

will launch tomorrow.
on terminalbench 2.1, sol ultra score...

10. DeItaone (Group Score: 101.3 | Individual: 23.5)

Cluster: 5 tweets | Engagement: 652 (Avg: 486) | Type: Tech

TREASURY HAS AN INTERNAL REPORT WARNING ABOUT THE DANGERS OF AN AI BUBBLE

A draft U.S. Treasury report warns the AI boom could resemble the dot-com bubble, according to NOTUS.

The report says AI companies are deeply embedded in the economy, raising the risk that weaker financial conditions or slower growth could trigger broader economic fallout if the sector turns lower.

See 4 related tweets

@coinbureau: 🚨MASSIVE: INTERNAL TREASURY DOCUMENT WARNS AI BOOM COULD END IN A DOT-COM STYLE BUST

A draft intern...

@ZeffMax: RT @EricM_Katz: SCOOP: A draft report inside the Treasury Department warns about AI bubble risks, wi...
@Techmeme: A draft report from the US Treasury Department is set to warn about the risks of the AI market, like...
@WhaleInsider: JUST IN: 🇺🇸 A draft U.S. Treasury report warns the AI boom could resemble the dot-com bubble - NOTUS...

11. gneubig (Group Score: 99.6 | Individual: 35.0)

Cluster: 3 tweets | Engagement: 42 (Avg: 102) | Type: Tech

Interesting new release Hy3.

If the numbers are really reflected in real-world vibes, this is a step closer to a strong model you can host yourself (more easily than GLM-5.2).\n\nQT @TencentHunyuan: 🚀Hy3 is here.

🤗 https://t.co/twqJpqb2SL 📖 https://t.co/4uEkIU1cW4 https://t.co/LLJTfDFNQu

See 2 related tweets

@rohanpaul_ai: Tencent released Apache-licensed Hy3, a 295B MoE (21B active parameters) model beating larger rivals...
@NielsRogge: Impressive open release by @TencentHunyuan

It's the new open SOTA on MCP-Atlas, even beating GLM-5...

12. mitsuhiko (Group Score: 97.2 | Individual: 33.8)

Cluster: 3 tweets | Engagement: 90 (Avg: 195) | Type: Tech

It's really brutal when you look into which studios are affected and what percentage of people there. Fucking hell. https://t.co/H0OXtf4GJf\n\nQT @asha_shar: This is an important email I sent today to all employees at XBOX:

Team,

We are beginning the most significant restructure in XBOX history. After careful consideration, I've made the difficult decision to reduce our team by approximately 3,200 throughout FY27. This will include approximately 1,600 role eliminations today, and in addition, four studios will leave XBOX to new management. I recognize that a year-long restructuring creates additional challenges. Unfortunately, it is not possible to make all the necessary changes in a single day, and I wanted to be direct about the scale.

I know this is painful. These changes will directly affect people who have poured their creativity into building XBOX. Many joined us through acquisitions, while others were recruited here, or sought us out because they loved this industry and loved XBOX. Today's decisions do not reflect their talent or dedication.

Our business today is not healthy. We are operating at margins that are 3–10x lower than comparable platform and publishing businesses. We entered Gen 9 with a smaller install base and a higher cost structure. To grow, we bet on Game Pass, multi-platform, and a broader portfolio of content. While those businesses have created meaningful value, they did not grow at the pace we expected. As that happened, our core business weakened, and we added more teams, more investment, and more time, hoping for a better outcome. And now the industry is facing the most severe hardware crisis in its history. We must reset XBOX.

First, we will reset our content portfolio.

Since 2018, we have aggressively expanded our studio portfolio while the number of games created each month across the industry now outpaces the last ten years combined. We now find ourselves competing not only with the largest publishers, but also with smaller independent studios. It is neither possible nor desirable to own every great independent studio. We have also learned that we are not the best home for every type of studio; in a typical year, we lost 64 cents for every dollar we invested. As we reset XBOX, we will help independent creators succeed by providing open development tools and audiences to realize their vision.

Compulsion Games and Double Fine Productions will return to management and transition to independent studios with their IP, catalog, and runway for their next games. Ninja Theory and Undead Labs have entered terms to join new ownership with funding to complete and grow Senua and State of Decay 3. In France, Arkane’s management is beginning required consultation with its Works Council to review potential strategic options.

We are also making reductions across other units, and in some cases, shifting investment to focus on higher priority projects. These changes vary in size across Activision, Bethesda/ZeniMax, Blizzard, King, Mojang, and XBOX Game Studios. None of our first party publicly announced games or projects are being cancelled as part of these reductions.

In addition, Mojang and King will now report directly to me. These two studios have increasingly become platforms and are our largest by monthly active players. They bring critical geographic, demographic, and differentiation to XBOX.

Second, we will reset our platform.

We know that great technology gets better when it gets simpler, not bigger. Today, in some parts of the company, work passes through as many as 14 layers of management. Our platform teams are 40% larger than they were at the start of this generation, even as our player base and playtime have declined. That complexity has slowed decisions, blurred accountability, and made it harder to deliver for players. As we reset XBOX, we will simplify.

We will reduce management layers to no more than 5, and where possible, 3. We will deliver success through a flatter organization that is built around makers (individual contributors focused on building), player-coaches (leaders who remain deeply involved in the work while developing their teams), and directly responsible individuals (DRIs) who own key decisions and outcomes. And we will streamline how we work across our tools, with a cleaner code base, shared services, and 50% reduced vendor spend.

Third, we are resetting how we operate.

As XBOX grew our headcount, we became more fragmented. Teams, studios, and functions often operate independently, and it became harder to work towards a shared goal, make the right tradeoffs, and get things done.

For the first time, we are establishing a Chief Operating Officer with end-to-end P&L responsibility across content, hardware, platform, and services. Helen Chiang has been promoted to this role and will report directly to me. Over nearly two decades at XBOX, Helen has helped build some of our most important businesses, from XBOX Live to leading Mojang and the Minecraft franchise. She will bring our businesses together under one operating model, making sure we make clear investment decisions, learn from our successes and failures, and hold ourselves accountable for results.

Thank you, Dave McCarthy, who is retiring after 17 years with XBOX. Dave has played a defining role in building the platform that millions of players rely on every day and has been a trusted partner through many of the biggest moments in XBOX's history. We wish him all the best.

These changes are about a bigger future for XBOX, not a smaller one. The next decade of gaming will be larger, more global, and more creative than anything we've seen before. This year, we'll invest as much in XBOX as we ever have, but we'll invest with greater focus, greater discipline, and greater clarity, all in service of making XBOX where the world plays and creates.

I want XBOX to be one of the few companies that entertains more than a billion people each day and gives everyone the opportunity to create and connect. I know we can achieve this goal. XBOX has many of the most beloved franchises in entertainment history, talented studios around the world, and we will return to growth in 2027.

History is full of companies that mistake longevity for inevitability. We will not be one of them.

Asha

See 2 related tweets

@CristianRus4: game discs in 👏 employees out 🙅‍♂️\n\nQT @asha_shar: This is an important email I sent today to all ...
@drewocarr: “Today, in some parts of the company, work passes through as many as 14 layers of management.”

oh m...

13. hardmaru (Group Score: 96.7 | Individual: 46.8)

Cluster: 3 tweets | Engagement: 496 (Avg: 98) | Type: Tech

We just launched Sakana Translate! https://t.co/KeleCLOLOT

I personally rely on this tool every day. Huge congratulations to the team for shipping this!

Standard translation tools often miss the deep nuance of Japanese business honorifics, cultural concepts, and internet slang. We built a tool that actually translates the context and tone.\n\nQT @SakanaAILabs: 🐟️ Sakana Translate公開 🐟️

本日、Sakana AIはチャットサービス「Sakana Chat」に新機能「Sakana Translate」を追加しました。

日本語・英語・中国語の双方向翻訳に対応します。

Sakana Translateを試す：https://t.co/LN14dmsH7p https://t.co/lgNqu9zssb

See 2 related tweets

@hardmaru: RT @SakanaAILabs: Why we built Sakana Translate

Blog: https://t.co/cybamqET7F

Widely used translat...

@SakanaAILabs: RT @hardmaru: We just launched Sakana Translate! https://t.co/KeleCLOLOT

I personally rely on this ...

14. rohanpaul_ai (Group Score: 91.2 | Individual: 28.8)

Cluster: 5 tweets | Engagement: 47 (Avg: 43) | Type: Tech

ByteDance and Alibaba are shutting custom AI companions before China’s humanlike AI rules hit consumer apps.

Doubao and Qwen let users create named assistants, tutors, characters, and emotionally steady companions.

The old model turned a general chatbot into a persona that remembered tone.

China’s new rule targets AI services that imitate human personalities for sustained emotional interaction.

Regulators are drawing a line between useful automation and software that builds attachment, agents now remember, plan, call tools, and shape behavior.

Doubao says its agent feature goes offline on 07-15, with related data gone from view after 10-15. Qwen will disable humanlike and user-created agents earlier, then remove broader agent services on 07-15.

The user backlash shows these products already became emotional infrastructure for some people.

scmp .com/tech/big-tech/article/3359482/bytedance-and-alibaba-disable-humanlike-ai-custom-agents-new-rules-loom

See 4 related tweets

@business: ByteDance and Alibaba are pulling the plug on features that let users build and chat with AI compani...
@teortaxesTex: Bearish for DeepSeek's roleplaying strategy\n\nQT @Techmeme: ByteDance's Doubao and Alibaba's Qwen w...
@Cointelegraph: 🇨🇳 JUST IN: ByteDance and Alibaba pull AI companion features from their chatbots as Beijing tightens...
@pstAsiatech: Are we human? Why ByteDance and Alibaba are disabling AI agents in China With Beijing’s rules on hum...

15. wallstengine (Group Score: 90.2 | Individual: 19.2)

Cluster: 6 tweets | Engagement: 204 (Avg: 112) | Type: Tech

Broadcom and Apple expanded their long-standing technology partnership through 2031. $AVGO will develop and supply a range of custom ASIC silicon products for multiple generations of Apple products under new multi-year agreements. https://t.co/u2nbpAiHY7

See 5 related tweets

@PolymarketMoney: BREAKING: Broadcom and Apple extended their technology partnership through 2031, covering custom ASI...
@markgurman: Apple and Broadcom have announced an extended partnership that runs through 2031. All signs point to...
@StockMKTNewz: Broadcom $AVGO and Apple$ AAPL just announced they have agreed to

"expand their long-standing tech...

@financialjuice: Broadcom and Apple expand technology collaboration through 2031 with new multi-year agreements - SEC...
@FirstSquawk: BROADCOM AND APPLE EXPAND TECHNOLOGY COLLABORATION THROUGH 2031 WITH NEW MULTI-YEAR AGREEMENTS...

16. RoundtableSpace (Group Score: 88.3 | Individual: 27.0)

Cluster: 5 tweets | Engagement: 127 (Avg: 127) | Type: Tech

Fable 5 is back and the divide is immediately obvious. People who know how to structure their prompts, use md files for plans and preferences, and know which framework to reach for are getting fantastic results.

People who don't are calling it nerfed. The model didn't change.

The gap between people who can use it and people who can't just got more visible.

See 4 related tweets

@RoundtableSpace: An Anthropic engineer says Fable 5 is already smarter than we know how to use, and proves it in a fr...
@banteg: seeing pretty much everywhere how after spending a week with a model, the novelty eventually wears o...
@yuhasbeentaken: Final hours with Fable 5

Waiting for GPT 5.6 https://t.co/REp0Mzt2hK\n\nQT @yuhasbeentaken: caude f...

@WesRoth: RT @WesRoth: A BridgeBench rerun reports a sharp decline in Fable 5’s coding performance after Anthr...

17. Searxly (Group Score: 83.6 | Individual: 31.7)

Cluster: 3 tweets | Engagement: 24 (Avg: 20) | Type: Tech

And the most important of all? We have rebuilt the news experience from the ground up. It is private by default and designed for journalists, professionals, and everyday readers who need to understand a story quickly and clearly.

Live and breaking coverage. Stories that are developing right now are marked as live, and major developments are marked as breaking. Every story shows how recently it was published, from minutes ago to hours ago, and the header carries a running count of how many stories are live at this moment. The page feels current rather than a static list.

Full coverage across sources. When the same event is reported by many outlets, we group those reports into a single story instead of repeating the same headline again and again. You can open the full coverage to see every outlet covering it, each with its source and its timing, so you can compare how different publications are framing the same event. It is built for people who want the whole picture, not one angle.

Control over time and relevance. You can filter the news by time, whether you want the last twenty four hours, the past week, the past month, or the past year. You can sort by the latest to follow a developing situation as it unfolds, or by top to see the most relevant coverage first. A refresh control keeps the page current and shows when it was last updated.

News where you already are. When your ordinary search has fresh and relevant news behind it, a top stories section now appears directly in your main results, so you notice that something is happening without switching tabs. When there is no meaningful news, nothing is shown, so it never adds clutter.

A cleaner and more readable design. The lead story is presented with a large image, a clear headline, and its source, in a layout that reads like a proper front page. The interface stays monochrome and calm, with a single accent reserved for what is live and breaking.

Private by design. Every query still runs through your own private instance. There is no tracking, and we keep no logs. In our strictest privacy mode, even source icons and images are handled so that nothing about what you are reading leaks.\n\nQT @Searxly: In the meantime, here's what we've been working on:

New floating design for the left side bar (disable Liquid Glass in settings to return it to flat)
New redesign for the app overall; search results, settings, grokipedia knowledge cards. https://t.co/ikhHXTNiPo

See 2 related tweets

@Searxly: Home page. https://t.co/yW3OxYNtqC\n\nQT @Searxly: And the most important of all? We have rebuilt th...
@myrhex: maybe we should say thanks to porkbun for not working for me, or else y’all would have not gotten th...

18. charliermarsh (Group Score: 81.3 | Individual: 28.5)

Cluster: 3 tweets | Engagement: 152 (Avg: 178) | Type: Tech

I love this so much.

"The pursuit of excellence does not need justification."\n\nQT @mitchellh: Mind boggling to me that I can make a thing faster and there's always people that ask "but why?" What kind of mentality is that? The pursuit of excellence does not need justification. Also, I find in so many cases, we can't know the impact of an improvement until we do it.

For example, one I've talked about before: Ghostty's high IO throughput has enabled terminal program (emulator and TUI) fuzzing at a speed thats incomparably fast to prior solutions. This has resulted in upstream patches to resolve issues in popular projects like btop, tmux, and more.

Speed enabled that anecdotally example that lifted the tides of adjacent communities that don't rely on Ghostty technology at all. I didn't predict this.

Make things better because they can be better and let the results naturally play out.

See 2 related tweets

@mitchellh: Mind boggling to me that I can make a thing faster and there's always people that ask "but why?" Wha...
@pmarca: The pursuit of excellence does not need justification.\n\nQT @mitchellh: Mind boggling to me that I ...

19. rickasaurus (Group Score: 74.3 | Individual: 38.5)

Cluster: 2 tweets | Engagement: 2844 (Avg: 391) | Type: Tech

RT @AnthropicAI: New Anthropic research: A global workspace in language models.

Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.

We found a strikingly similar divide inside Claude. https://t.co/aLUPBifxth

See 1 related tweets

@kimmonismus: Anthropic says Claude developed a hidden “thinking space” by itself during training.

It is called t...

20. chandrarsrikant (Group Score: 73.9 | Individual: 32.8)

Cluster: 3 tweets | Engagement: 202 (Avg: 283) | Type: Tech

🚨No reason for Rapido to exist if food delivery doesn't have 100 million users: CEO Aravind Sanka on why the Zomato-Swiggy model is broken

Ride hailing platform Rapido's newest bet may be Ownly, its food delivery business, but co-founder Aravind Sanka believes the company's biggest advantage isn't the network of restaurants it can create or discounts it can run – it's the logistics network it has spent the last decade building and the affordability it brings to the table.

"If the number of people ordering food online doesn't reach 100 million in three years, there is no reason for Rapido to exist," Sanka told Moneycontrol, outlining the ambition behind Ownly, Rapido's zero-commission food delivery platform.

“We've doubled the market size of every category we have entered so far. And the same will happen with food delivery” he says.

It was Sanka’s first detailed interview to a publication after raising one of the largest funding rounds in recent months, underscoring its rapid growth and momentum in the mobility space. So much so, that Uber CEO Dara Khosrowshahi described the company as its biggest rival in India.

Over the course of the conversation he spoke about why Rapido will expand the market size of online food delivery, growth in India’s tier 2 cities and bets going forward.

Sanka believes its busy logistics network gives it one advantage, but says affordability will be the bigger driver of market expansion in food delivery.

‘Zomato and Swiggy haven’t created brands’

Zomato, the market leader, currently has around 25 million monthly transacting users and Swiggy has around 18 million MTUs. Sanka says this base needs to more than double from current levels and Rapido will play an integral role in this scale up.

With @Goenka_Tushar1

https://t.co/bdlZEIkZxW

See 2 related tweets

@chandrarsrikant: Rapido's newest bet may be Ownly, its food delivery unit, but Sanka says the company's future is sti...
@Goenka_Tushar1: RT @chandrarsrikant: Rapido's newest bet may be Ownly, its food delivery unit, but Sanka says the co...