科技推文精选 - 2026年6月29日

2026年6月29日科技每日简报

Today's top tech conversations are led by @martin_casado, whose post about 'RT @elonmusk: Grok 4.5, based ...' garnered the highest engagement. Key themes trending across the top stories include model, models, https, usage, codex. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. martin_casado (Group Score: 573.4 | Individual: 46.1)

Cluster: 26 tweets | Engagement: 7732 (Avg: 601) | Type: Tech

RT @elonmusk: Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.

RL is continuing to significantly improve the model, and the Grok Build harness gets better every day.

Nice work by all those involved!

Completely trained from scratch new models will be released by @SpaceX every month this year.

See 25 related tweets

@scaling01: not only is Grok 4.5 coming, but this year SpaceX is going to release a model EVERY month

(in Elon ...

@teslaownersSV: BREAKING: Elon Musk says Grok 4.5 is now in private beta at SpaceX and Tesla.

Built on the 1.5T V9 ...

@kimmonismus: I hope Elon is referring to opus 4.8. would be a welcome release if grok 4.5 outperforms opus 4.8 on...
@FirstSquawk: Elon Musk: Grok 4.5, built on 1.5T V9 foundation model with cursor data in extra training, now in pr...
@Hesamation: "come back when you're blocked by the government"

Elon could aurafarm so hard by just open sourcing...

2. StockMKTNewz (Group Score: 179.3 | Individual: 30.5)

Cluster: 10 tweets | Engagement: 557 (Avg: 345) | Type: Tech

Google $GOOGL has reportedly placed limits on Meta Platforms$ META use of its Gemini AI models due to computing capacity constraints

The restrictions have affected Meta's internal projects and the company has told staff to make more efficient use of AI tokens - Financial Times https://t.co/zLKhB4mhrB

See 9 related tweets

@business: Google has placed limits on Meta’s use of its Gemini AI models because it could not provide as much ...
@rohanpaul_ai: FT: Google capped Meta’s use of Gemini after Meta asked for more model compute capacity than Google ...
@FT: Google has put limits on Meta’s use of its Gemini AI models after the social media giant sought more...
@wallstengine: $GOOGL has reportedly limited Meta’s use of Gemini after$ META sought more AI compute capacity than ...
@jukan05: Google reportedly limited Meta’s use of Gemini due to a shortage of compute resources. — FT

Google ...

3. FirstSquawk (Group Score: 151.0 | Individual: 40.9)

Cluster: 7 tweets | Engagement: 474 (Avg: 64) | Type: Tech

ZHIPU AI’S NEW MODEL REPORTEDLY MATCHES CLAUDE MYTHOS IN FINDING SECURITY VULNERABILITIES

A new AI model from Zhipu AI reportedly delivers security bug detection performance comparable to Anthropic’s Claude Mythos. The development underscores intensifying competition in advanced AI, particularly in cybersecurity and automated vulnerability research.

See 6 related tweets

@VadimStrizheus: Dario watching a Chinese lab make a open-sourced Mythos model https://t.co/yVGJ7u7i71\n\nQT @Polymar...
@rauchg: Mythos / Sol cybersecurity capabilities are equally useful in an offensive as well a defensive capac...
@EHuanglu: china is going to open source mytho https://t.co/0gZFHe81pb\n\nQT @Polymarket: JUST IN: A new Chines...
@DataChaz: US Frontier labs right now: https://t.co/a106qyQZFE\n\nQT @Polymarket: JUST IN: A new Chinese AI mod...
@peterwildeford: This is fake news lol\n\nQT @Polymarket: JUST IN: A new Chinese AI model from Zhipu AI reportedly ma...

4. matei_zaharia (Group Score: 146.3 | Individual: 48.0)

Cluster: 4 tweets | Engagement: 224 (Avg: 50) | Type: Tech

Good lessons on managing AI spend. We support a lot of this natively on Databricks with Unity AI Gateway, which makes it easy to analyze and control usage in one place, and it’s also easy to set these up as policies with the open source https://t.co/zh1P01h1B5 framework.\n\nQT @brian_armstrong: How to keep AI spend flat while token usage grows exponentially: Not with friction and spend alerts. With better defaults, routing, and caching.

Better Defaults (not Usage Caps) – Engineers can choose any model they want, but defaults matter. We’re experimenting with defaulting to open weight models like GLM 5.2 and Kimi 2.7 through our LLM gateway, while still encouraging engineers to choose the right model for the task. 91% of our employees were never hitting their usage caps, so instead of lowering caps and driving up alerts, we're moving to cheaper defaults. Note that code reviews use a diversity of models, so they can check each other's work.

Better Routing – In our custom harnesses, we preprocess prompts and route to the best model for the job, considering cache hits and model pricing. For instance, you may want a frontier model for planning, but not for execution where they can be overkill. Ultimately, humans shouldn't be choosing models - AI can automate this task.

Better Caching – Cache misses are the easiest way to drive your cost up. All of our requests are cache aware, so we’re reusing a warm cache wherever possible. For example, our cache hit rate went from 5% → 60% in LibreChat once properly implemented.

Keep Context Lean – Start fresh sessions when switching tasks. Scope file context narrowly. Disconnect unused tools. Don't just compact. The goal isn't fewer tokens used, it's fewer tokens wasted.

Better Visibility – Our engineers can use as many tokens as they want, from whatever model they want, but we’ve made usage visible – and the more you spend on AI, the more impact we expect.

The goal isn't to suppress usage. It's to build the infrastructure that makes exponential growth sustainable.

Putting this into practice has cut our AI spend nearly in half, while our token usage continues to grow.

See 3 related tweets

@ClementDelangue: the future of AI is multi-model (including a majority of open-source ones provided by @huggingface o...
@cryptopunk7213: few things that’ll become increasingly more valuable over the next 6 months:
which ai models you ...
@bradmenezes: Claude fever is over.

This is the end for frontier lab AI apps.

Open source will eat 80% of all ...

5. mark_k (Group Score: 124.6 | Individual: 40.5)

Cluster: 4 tweets | Engagement: 282 (Avg: 192) | Type: Tech

Former White House AI czar David Sacks is no fan of Washington's recent push to restrict access to frontier AI models. His argument is simple: America does not win the AI race by slowing itself down. It wins by building faster, scaling infrastructure, lowering energy costs, and exporting its technology before China fills the gap.

The WSJ report he cites is exactly the nightmare scenario: Chinese models catching up in sensitive areas like cybersecurity while U.S. policy debates how much of its own lead it should voluntarily restrain.

Restricting American AI may feel like safety. But if it accelerates Chinese alternatives, it is not safety. It is surrender with extra paperwork.\n\nQT @DavidSacks: A year ago, President Trump declared that America was in a global AI race and that the way to win it was to be pro-innovation, pro-infrastructure, pro-energy, and pro-export. President Trump was exactly right; we deviate from that strategy at our peril. https://t.co/XXpesCPMBp

See 3 related tweets

@BullTheoryio: 🚨 CHINA JUST MATCHED ANTHROPIC'S TOP MODEL IN CYBERSECURITY.

And It's A Big Problem For Us Ai Stoc...

@chiefofautism: I don’t get how Nick Land-coded e/acc people can cheer boxed superintelligence into existence, then ...
@bradmenezes: China is closing the gap while the US government is slowing down model releases.

This market dyna...

6. nathanbenaich (Group Score: 121.7 | Individual: 26.2)

Cluster: 6 tweets | Engagement: 81 (Avg: 20) | Type: Tech

i’m not sure that’s how this works…\n\nQT @business: Austria is pushing the EU to consider hosting Anthropic within its borders to counter US efforts to block foreigners from using its most advanced AI models. https://t.co/e9mn1EAQBx

See 5 related tweets

@business: Austria is pushing the EU to consider hosting Anthropic within its borders to counter US efforts to ...
@wallstengine: Austria is pushing the EU to consider hosting Anthropic inside Europe after the U.S. restricted fore...
@Reuters: Austria lobbies EU to host Anthropic AI after US curbs, Bloomberg News reports https://t.co/1KtXq8GM...
@Reuters: Austria urges Europe to host Anthropic following US curbs on AI access https://t.co/26seiMetZt https...
@Techmeme: Letter: Austria is pushing the EU to consider hosting Anthropic within its borders, highlighting EU ...

7. business (Group Score: 110.1 | Individual: 29.3)

Cluster: 5 tweets | Engagement: 113 (Avg: 84) | Type: Tech

The bill for the AI era has officially come due for Apple customers, with the company dramatically raising prices just as it was trying to make products like the Mac more affordable, writes Mark Gurman.

Read the latest edition of the Power On newsletter: https://t.co/Y3oZuMofso

📷️: Qilai Shen/Bloomberg

See 4 related tweets

@markgurman: There’s no question that ChatGPT has created value. But the costs are now becoming real to many cust...
@markgurman: The irony is that Apple never wanted this version of the AI era in the first place. Apple has mostly...
@markgurman: Power On: Apple’s sweeping price increases bring home the costs of the AI era for many people for th...
@business: The AI boom made Apple’s price increases on laptops and other products inevitable https://t.co/qodvw...

8. alexocheema (Group Score: 106.3 | Individual: 48.5)

Cluster: 4 tweets | Engagement: 988 (Avg: 167) | Type: Tech

We’ve been working with NVIDIA in their HQ for the past month.

We’re going to make Local AI The Default.

BIG news to share at Local AI summit, SF, July 2nd. https://t.co/K8qVfLEcwo\n\nQT @TheAhmadOsman: MASSIVE NEWS

Teamed up with NVIDIA to make Local AI The Default https://t.co/kmGgcBEZ4f

See 3 related tweets

@aiDotEngineer: RT @TheAhmadOsman: MASSIVE NEWS

Teamed up with NVIDIA to make Local AI The Default https://t.co/kmG...

@ivanfioravanti: RT @alexocheema: We’ve been working with NVIDIA in their HQ for the past month.

We’re going to make...

@alexocheema: RT @msharmavikram: Incredible work is happening by Team Exo and @nvidia over the last few weeks and ...

9. danshipper (Group Score: 102.9 | Individual: 36.9)

Cluster: 3 tweets | Engagement: 30 (Avg: 13) | Type: Tech

self recommending\n\nQT @lennysan: Andrew Ambrosino (@ajambrosino) leads the team behind the Codex desktop app at @OpenAI. Codex usage has 6x'd since February, reaching over 5M weekly active users, and nearly 100% of OpenAI's employees use the Codex app regularly (and not just the engineers).

Andrew's personal mission is to build "the best desktop app that has ever existed, full stop." If you've used the Codex app lately, you know he's not far off from that goal.

In our in-depth conversation, we discuss: 🔸 The "zone defense" model of how PMs at OpenAI operate 🔸 Why AI is so bad at design 🔸 Why Andrew thinks the Codex app would have flopped if they'd shipped it in November instead of February (same product—only the model changed) 🔸 What “taste” really means as a professional skill 🔸 How Andrew uses Codex to run his workflows 🔸 His vision for Codex + ChatGPT

Listen now 👇 https://t.co/rVoohRbCiu

See 2 related tweets

@lennysan: "When anybody can build anything, the taste to know what to build becomes the whole game." — @ajambr...
@pvncher: RT @lennysan: Andrew Ambrosino (@ajambrosino) leads the team behind the Codex desktop app at @OpenAI...

10. FirstSquawk (Group Score: 99.1 | Individual: 50.0)

Cluster: 3 tweets | Engagement: 441 (Avg: 64) | Type: Tech

LAWMAKERS REPORTEDLY LEFT SHAKEN AFTER CLOSED-DOOR MYTHOS AI DEMO SHOWED POTENTIAL FOR SOPHISTICATED CYBER AND CRIMINAL MISUSE

According to reports, a private demonstration of Anthropic’s Mythos AI model left some U.S. House members deeply concerned after it allegedly showcased capabilities that could be exploited to facilitate financial theft, such as draining bank accounts, and assist in planning serious crimes. The reported demonstration heightened concerns about the security risks posed by increasingly capable AI systems.

See 2 related tweets

@BullTheoryio: 🚨AI Emptied Bank Accounts in Minutes and Planned a Lawmaker's Kidnapping in 30 Seconds

House Homela...

@chiefofautism: RT @WhaleInsider: JUST IN: 🇺🇸 U.S. Lawmakers reportedly left shaken after closed-door Mythos AI demo...

11. shanaka86 (Group Score: 92.0 | Individual: 33.8)

Cluster: 3 tweets | Engagement: 62 (Avg: 133) | Type: Tech

A European government just asked the EU to give an American AI company (Anthropic Claude in this case) a new home, and the offer cannot land, for a reason that explains the entire new geopolitics of intelligence. Austria's letter to Brussels, urging Europe to host Anthropic after the US export-controlled its top models, is not a serious bid. It is a confession that Europe has realized it no longer controls the most important technology of the century.

The facts are real and narrow. Austria's digitalization chief wrote to the European Commission proposing the bloc explore the strategic establishment of Anthropic in the EU, offering legal certainty, capital, and shared values. Reuters could not independently verify the underlying report. Anthropic did not respond. And the author himself admitted there would be skepticism about whether it is even possible, while offering no plan, no funding figure, no infrastructure. A wish, dressed as a proposal.

The wish dies on physics. A frontier model is not a headquarters you relocate for a tax break. It is welded to hundreds of thousands of the newest chips, to gigawatts of power, to data centers signed to American hyperscalers years in advance. The AI sector is on track to need roughly 50 gigawatts of new US electricity by 2028, twice the peak draw of New York City, and Anthropic is spending tens of billions building exactly that, inside America. You can move a logo across an ocean. You cannot move the power plants.

Notice the deeper turn. For thirty years, nations competed to attract the company. Austria is trying to attract the mind. And it cannot, because this mind now runs on an industrial base only America presently commands, the same base that let Washington switch the models off in the first place.

Whoever owns the sockets and the chips owns who gets to think at frontier speed. Europe is offering an address to something chained to a building it cannot leave. yes!\n\nQT @coinbureau: 🚨AUSTRIA URGES EU TO HOST ANTHROPIC AFTER THE US SHUT DOWN ITS MOST ADVANCED AI FOR THE ENTIRE WORLD

Austria called for the "strategic establishment of Anthropic within the European Union" in a letter to the European Commission per Bloomberg.

The move comes after the US imposed export controls on Anthropic's Fable 5 and Mythos 5 on June 12, blocking all foreign nationals from accessing them.

Anthropic was forced to disable both models worldwide because it could not verify user nationality in its cloud systems.

See 2 related tweets

@rohanpaul_ai: CEO of Box, @levie on AI access as the new blunt instrument in geopolitics.

“Imagine AI being used...

@kimmonismus: Austria is trying to lure Anthropic to Europe.

But it seems more like an act of desperation. After ...

12. Parul_Gautam7 (Group Score: 91.7 | Individual: 28.4)

Cluster: 4 tweets | Engagement: 214 (Avg: 110) | Type: Tech

I like seeing projects that rethink the problem instead of just scaling the model.

BrowserBC looks worth keeping an eye on.\n\nQT @vida_agent: We open-sourced BrowserBC: A system that turns human browser trajectories into reusable agent skills. Just one recording is enough to generalize a skill.

🛠️ GitHub: [https://t.co/WP8mQGuJ6N]

Here’s how it works. 👇

See 3 related tweets

@svpino: This is all open source, and it's probably the direction we'll be automating a ton of tasks in the f...
@FellMentKE: Most agent moats are just companies spending more on compute.

But the scaling story I actually buy ...

@FellMentKE: RT @vida_agent: We open-sourced BrowserBC: A system that turns human browser trajectories into reus...

13. AnatoliKopadze (Group Score: 89.3 | Individual: 27.3)

Cluster: 4 tweets | Engagement: 148 (Avg: 870) | Type: Tech

This prompting guide from Anthropic engineers is something absolutely everyone should watch.

24 minutes of the fundamentals you need whether you build AI agents, loops, or anything else.

Prompting is what every other AI skill is built on and almost no one learned it the right way.

Watch it, then read the guide on everything Claude can do below.\n\nQT @AnatoliKopadze: https://t.co/AAWIZD1pNL

See 3 related tweets

@mikenevermiss: anthropic research lead:

“99% of our engineers use 300+ AI agents working together.”

“close the lo...

@petergyang: RT @petergyang: The best AI teams are moving from prompting AI to building agents that can work over...
@DataChaz: RT @DataChaz: STOP PAYING $500 FOR AI COURSES

If you want to build your first AI agent, Google just...

14. eng_khairallah1 (Group Score: 88.2 | Individual: 59.1)

Cluster: 2 tweets | Engagement: 1068 (Avg: 113) | Type: Tech

GOOGLE CEO SUNDAR PICHAI: "IF YOU DON'T LEARN HOW TO ORCHESTRATE AGENTS NOW, YOU'LL SPEND 2027 CATCHING UP TO PEOPLE WHO STARTED TODAY."

30 minutes on why the best engineers stopped writing code line by line and started orchestrating agents instead.

Most people think building an agent requires an engineering degree.

It doesn't.

It requires one guide and one afternoon.

Watch the interview. Then read the article below.

One guide. One afternoon. That's all it takes.\n\nQT @eng_khairallah1: https://t.co/mLfo3ZV2Vw

See 1 related tweets

@eng_khairallah1: RT @eng_khairallah1: GOOGLE CEO SUNDAR PICHAI: "IF YOU DON'T LEARN HOW TO ORCHESTRATE AGENTS NOW, YO...

15. diptanu (Group Score: 87.8 | Individual: 34.0)

Cluster: 3 tweets | Engagement: 28 (Avg: 21) | Type: Tech

When a business depends 100% on a model, using an open weight model on multiple providers is a lot better strategy than relying on the latest closed frontier model. People dont necessarily need to be inference experts. There is Baseten, Fireworks, Fal and countless other companies figuring out how to make inference fast on open weights models.\n\nQT @rohanpaul_ai: Anthropic CEO Dario Amodei on Open-Source AI Models.

"I don't think open source works the same way in AI that it has worked in other areas. Primarily because with open source you can see the source code of the model. Here we can't see inside the model, it's often called open weights instead of open source to kind of distinguish that. But a lot of the benefits, which is that many people can work on it and that it's kind of additive, don't quite work in the same way.

So I've actually always seen it as a red herring. When I see a new model come out I don't care whether it's open source or not. If we talk about Deep Seek I don't think it mattered that Deep Seek is open source. I think I ask, is it a good model? Is it better than us at the things that matter? That's the only thing that I care about.

It actually doesn't matter either way. Because ultimately you have to host it on the cloud. The people who host it on the cloud do inference. These are big models, they're hard to do inference on.

When I think about competition I think about which models are good at the tasks that we do. I think open source is actually a red herring.

It's not free. You have to run it on inference and someone has to make it fast on inference."

From 'Alex Kantrowitz' YT channel (full video link in comment)

See 2 related tweets

@AskYoshik: If Dario is saying "don't use open-source AI models"

then you should definitely use open-source.\n...

@rohanpaul_ai: RT @rohanpaul_ai: Anthropic CEO Dario Amodei on Open-Source AI Models.

"I don't think open source ...

16. chiefofautism (Group Score: 86.1 | Individual: 32.2)

Cluster: 3 tweets | Engagement: 18 (Avg: 297) | Type: Tech

you forgot to also mention that half of sillicon valley building & running on chinese models

digging own grave because its cheaper is so american https://t.co/n2VkJjsUYB\n\nQT @ihtesham2005: China just released an open source AI model that matches the best closed models from OpenAI and Anthropic. Gavin Baker explained exactly how they did it and the answer should concern every American AI lab.

The model is called GLM 5.2. It was built by Z. AI.

You get 744 billion parameters, 1 million token context window and its MIT license, meaning anyone can download it, fork it, build a company on it, with no restrictions and no Dario.

It scored 51 points on the artificial analysis intelligence index. The highest score any open weight model has ever achieved.

It beat GPT 5.5 on the frontier software engineering benchmark. It trails Claude Opus 4.8 by less than one percentage point. And it costs 85% less to run than GPT 5.5 for comparable performance.

Gavin Baker said on the All-In podcast that this model has challenged some of his beliefs. Then he explained how China built it.

The method is called distillation.

Just think of tens of thousands of phones and computers running simultaneously, all hitting the frontier model APIs through masked accounts, asking specific questions, and harvesting what happens inside the model when it answers. Every reasoning step, every token. The entire thinking process gets recorded and fed back into the Chinese model during training.

It is a cheat sheet. It is the answer key to the exam.

And here is the part that should worry everyone.

Sacks said it plainly. China was already nine months behind American models. But now that GLM 5.2 is good enough to run its own reinforcement learning, it can improve itself without needing to distill from American models anymore. The cheat sheet let them get close enough to start writing their own answers.

Sacks said we are six months behind on the model and 24 months behind on silicon and they are only a few months behind in total.

The Z. AI founder told Elon Musk directly that open weight fable-level capability will be here before Q1 2027.

Every restriction Anthropic lobbied for, every self-imposed safety guardrail, every month of delay in releasing American frontier models accelerated this. The Chinese labs were not under those restrictions. They were not going to wait.

The composable model future Gavin described, where every enterprise runs a frontier model alongside their own fine-tuned open weight model, is coming regardless of what American labs do next.

The question is just whether the open weight half of that stack is American or Chinese.

Right now it is Chinese.

WATCH THE FULL PODCAST ON @theallinpod

See 2 related tweets

@negligible_cap: China might very well be winning the AI race considering intelligence created per dollar of capex sp...
@random_walker: RT @sayashk: Many people have written about advanced AI being too restricted, but my fear is we have...

17. ShanuMathew93 (Group Score: 83.2 | Individual: 32.8)

Cluster: 3 tweets | Engagement: 13 (Avg: 38) | Type: Tech

Great thoughts. Even if we get through this noise this time, the precedent has been set.

Andrew simply lays out what is likely to occur in the future and I get nervous around two specific points:

governments will require active access into the development efforts of the frontier labs because now that the tipping point has been crossed there’s always one new development away from the next vulnerability creating model. Don’t really love the idea of governments directly overseeing company efforts as that introduces a host of scary prospects around government power over citizens and businesses
frontier access between countries - allies or enemies is about to get really weird. Moving forward there will be a point where countries restrict access based ok nationality given the risk and that has ramifications for global coordination, where people live or businesses are based, and so forth\n\nQT @AndrewCurran_: I think both Fable 5 and GPT-5.6 get approved for general release next week, and for use outside of the United States as well. But people should remember this moment and remember this feeling, because it is almost inevitable that we eventually reach a point where approval does not arrive.

Capitalism is going to tip the scales this time. I doubt they will approve one model and not the other, because doing so would be seen as incredibly anti-competitive. Fable and GPT-5.6 will probably receive the same clearance, probably on the same day. I also doubt they want to restrict sales outside the US, because that would be seen as anti-business and would trigger a major backlash against American closed-source AI. The rumblings of which you can already hear today. There is also a plan now taking shape on both the US left and right to create some version of an AI public wealth fund that pays a dividend directly to American citizens. That fund needs to be fed by the global sale of the big labs top models to people outside the US. So I think there will be no freeze on their use outside the United States this time.

The other reason is that allowing this will make people happy, and it will soften the fact that Mythos, as was announced yesterday, is available only to a vetted group of US agencies and companies. I do not think that this basic structure will change from here on out. Mythos may eventually be made available to certain allies, but only after the US government, its agencies, and then some chosen American companies have access to Mythos-2, Sol-2, or whatever the new uber-model turns out to be.

I do not think this gap ever closes again, not even for allies. And that means the US will increasingly possess an intelligence advantage that touches almost everything: voting, markets, corporations, academia, infrastructure, and the internal operations of foreign states. Having Mythos-n will always be trumped by whoever has Mythos-n+1. Anthropic themselves have said within nine months Mythos will look like a toy. That advantage, standing at the top of this tower, is too large to give up voluntarily. It also means that many things will become suspect. People will see shadows everywhere. Barring espionage, a deliberate leak, or the emergence of a non-US competitor at the top end of the scale, this structure will persist for some time. The public fight is about access to models. But the real fight is about access to the future. And from this point forward, whoever holds this power will also become increasingly capable of keeping it for themselves.

See 2 related tweets

@inafried: RT @AndrewCurran_: I think both Fable 5 and GPT-5.6 get approved for general release next week, and ...
@alex_prompter: RT @alex_prompter: Anthropic's Fable lasted three days, Mythos never went public at all, and now Ope...

18. pmarca (Group Score: 82.2 | Individual: 30.1)

Cluster: 3 tweets | Engagement: 568 (Avg: 766) | Type: Tech

Interesting.\n\nQT @curiouswavefn: As a scientist, AI has made me feel the most intellectually alive and excited I have felt since I was a graduate student and postdoc more than 20 years ago. Every day I can start with an idea in the morning, and by lunchtime, I see a testable, rational, well-thought-out hypothesis forming in front of my eyes. And every day, the possibilities seem endless, like mountains beyond mountains. What a time to be alive.

Here's a case in point. I'm collaborating with a professor, an experimentalist, who is trying to solve a thorny problem in his field. There's one particular molecule that he is using in his experiments that seems to result in radically different crystal structures compared to similar molecules. What's happening here? He has come up with a few different hypotheses that could explain the differences but is not a theoretician and needs to tease them apart.

On Thursday, I started an investigation using AI at his bequest. The AI immediately confirmed the hypotheses that he had in mind and added a few of its own.

Then it started its exploration. The investigation was carried out in three different phases, each of increasing difficulty; the first one using classical physics, and the second and third using quantum mechanical techniques of increasing rigor. This tiered strategy is the right one.

By Thursday evening, I had the glimpse of an answer. Most of the hypotheses had been examined and rejected. Two stood out, although the AI identified one as more a mechanism through which the other one operated rather than a root cause. It immediately pivoted to the higher-level, more rigorous calculation.

Every time I interacted with the AI, it was more like a dialogue between a professor and a bright student or scientific collaborator than a mandate issued to a tool. The feeling was very much of a process where the AI and I were solving a problem together. I steered the conversation several times, pushed back, suggested course-corrections, acknowledged my own wrong ideas as well as the AI's and went back and forth. The AI was successful in keeping multiple requests in its memory, stacking them by priority while never losing the conversation thread.

By late Friday morning, there had collected enough data from the more rigorous calculation to corroborate the suspicion that it was really just one hypothesis that was the root cause. It then moved on to the next step, which was to come up with a distinct set of novel molecules that would confirm the hypothesis beyond any reasonable doubt. In addition, it launched an even more rigorous calculation at a higher level of theory.

By the end of Friday, roughly 48 hours later, using this multi-layered approach of increasing rigor, backed up by references, and made useful and actionable by testable experiments, the AI had arrived at a solid, rigorous conclusion.

Now imagine doing this every day, about any topic under the scientific sun, in any scientific field, so that your intellectual labor is multiplied a million-fold.

Mountains beyond mountains. What a time to be alive.

See 2 related tweets

@prz_chojecki: Research has never been so fun as it is right now. Doing science is now limited by your imagination ...
@pmarca: RT @curiouswavefn: As a scientist, AI has made me feel the most intellectually alive and excited I h...

19. narendramodi (Group Score: 81.8 | Individual: 27.9)

Cluster: 3 tweets | Engagement: 3613 (Avg: 10407) | Type: Tech

Had an excellent meeting with Dr. Navinchandra Ramgoolam, the Prime Minister of Mauritius. We met at the start of the year during the AI Summit and now we have the chance to meet in Seychelles.

We discussed ways to strengthen the India-Mauritius partnership. The progress being made across every area of our Enhanced Strategic Partnership is a matter of satisfaction for us. A number of projects are being implemented under our Special Economic Package.

We also discussed the way ahead with regard to sectors such as capacity building, skilling, defence, energy, cyber security and more.

@Ramgoolam_Dr

See 2 related tweets

@ANI: #WATCH | Victoria | During the press statement with Seychelles President Dr Patrick Herminie, PM Mod...
@ANI: #WATCH | Victoria | During the press statement with Seychelles President Dr Patrick Herminie, PM Mod...

20. Dan_Jeffries1 (Group Score: 77.7 | Individual: 31.7)

Cluster: 3 tweets | Engagement: 89 (Avg: 315) | Type: Tech

We desperately need an American open source champion to counter the tide of closed source companies ginning up fear in our politicians to capture the market.

I've got high hopes for these smart folks.

On a long enough timeline open will sweep away the darkness.\n\nQT @reflection_ai: Today we're sharing the next phase of Reflection.

We're building frontier open intelligence accessible to all.

We've assembled an extraordinary AI team, built a frontier LLM training stack, and raised $2 billion.

Why Open Intelligence Matters

Technological and scientific progress is driven by values of openness and collaboration.

The internet, Linux, and the protocols and standards that underpin modern computing are all open. This isn't a coincidence. Open software is what gets forked, customized, and embedded into systems worldwide. It's what universities teach, what startups build on, what enterprises deploy.

Open science enables others to learn from the results, be inspired by them, interrogate them, and build upon them in order to push the frontier of human knowledge and scientific advancement. AI got to where it is today through scaling ideas (e.g. self-attention, next token prediction, reinforcement learning) that were shared and published openly.

Now AI is becoming the technology layer that everything else runs on top of. The systems that accelerate scientific research, enhance education, optimize energy usage, supercharge medical diagnoses, and run supply chains will all be built on AI infrastructure.

But the frontier is currently concentrated in closed labs. If this continues, a handful of entities will control the capital, compute, and talent required to build AI, creating a runaway dynamic that locks everyone else out. There's a narrow window to change this trajectory. We need to build open models so capable that they become the obvious choice for users and developers worldwide, ensuring the foundation of intelligence remains open and accessible rather than controlled by a few.

What We've Built

Over the last year, we've been preparing for this mission.

We’ve assembled a team who have pioneered breakthroughs including PaLM, Gemini, AlphaGo, AlphaCode, AlphaProof, and contributed to ChatGPT and Character AI, among many others.

We built something once thought possible only inside the world’s top labs: a large-scale LLM and reinforcement learning platform capable of training massive Mixture-of-Experts (MoEs) models at frontier scale. We saw the effectiveness of our approach first-hand when we applied it to the critical domain of autonomous coding. With this milestone unlocked, we're now bringing these methods to general agentic reasoning.

We've raised significant capital and identified a scalable commercial model that aligns with our open intelligence strategy, ensuring we can continue building and releasing frontier models sustainably. We are now scaling up to build open models that bring together large-scale pretraining and advanced reinforcement learning from the ground up.

Safety and Responsibility

Open intelligence also changes how we think about safety. It enables the broader community to participate in safety research and discourse, rather than leaving critical decisions to a few closed labs. Transparency allows independent researchers to identify risks, develop mitigations, and hold systems accountable in ways that closed development cannot.

But openness also requires confronting the challenges of capable models being widely accessible. We're investing in evaluations to assess capabilities and risks before release, security research to protect against misuse, and responsible deployment standards. We believe the answer to AI safety is not “security through obscurity” but rigorous science conducted in the open, where the global research community can contribute to solutions rather than a handful of companies making decisions behind closed doors.

Join Us

There is a window of opportunity today to build frontier open intelligence, but it is closing and this may be the last. If this mission resonates, join us.

See 2 related tweets

@kimmonismus: The open internet was the training ground for modern AI.

Billions of texts, forum posts, code repos...

@_simonsmith: Yes, I hold two seemingly contradictory positions:

I think models and harnesses from OpenAI and ...