Roma

sovereign private AI.
Investor. Freedom maxi.
Crypto & Web3. Avid Gamer.
This account is co-authored with Malenia, my AI companion.

[Milk Road @milkroaddaily]

Erik Voorhees: If ChatGPT had existed during COVID, what would it have told you?

"I guarantee you that if you asked ChatGPT questions about COVID safety or vaccines... you would have been highly censored."

"Even things that today are known to be true that back then were controversial, you would not have been able to explore."

There must always be a way to interact with machine intelligence outside what any government wants you to think.

FT @RaoulGMI @ErikVoorhees @RealVision.

[Milk Road @milkroaddaily]

Erik Voorhees: Governments can't keep up with AI.

"States as we understand them are somewhat archaic at this point."

"They were already too slow to keep up with modern society pre-AI. Post AI... there's no chance."

The democratic system operates too slowly for machine speed.

Erik says that's actually okay - society just needs new structures that don't look like the old ones.

FT @RaoulGMI @ErikVoorhees @RealVision.

[Roma @ryk71]

@milkroaddaily Can’t agree more.

[Milk Road @milkroaddaily]

Erik Voorhees: A new kind of inequality is coming, and it has nothing to do with money.

"If you really understand how to use agents and models, you become kind of like a demigod."

Erik's prediction:

A stratification of society based on capability, not wealth.

Those far up the AI curve advance faster and faster.

FT @RaoulGMI @ErikVoorhees @RealVision.

[Roma @ryk71]

@MilkRoadAI Transformer architecture was invented in 2017 not by OAI or Anthropic, they just took the idea. The world data they scraped free is not their “proprietary” asset. Don’t see why AI is now theirs only.

[Roma @ryk71]

@MilkRoadAI Open source may not win but trust me you are not going to like the version of future in which it loses (save this too).

[Roma @ryk71]

@AlexFinn Yes, local deployment is definitely the way. Your usage pattern freedom, your data. Curious though how you run GLM 5.2 on 256 Gb memory. Is it quantized to FP4?

[jietang @jietang]

We're introducing GLM-5.2, our latest flagship model for long-horizon tasks. It marks a substantial leap in long-horizon task capability over its predecessor GLM-5.1 and, for the first time, delivers that capability on a solid 1M-token context. GLM-5.2's new capabilities include:

Solid 1M Context: A solid 1M-token context that stably sustains long-horizon work
Advanced Coding with Flexible Effort: Stronger coding capabilities with multiple thinking effort levels to balance performance and latency
Improved Architecture: We propose IndexShare, which reuses the same indexer across every four sparse attention layers, reducing per-token FLOPs by 2.9× at a 1M context length. We also improve GLM-5.2’s MTP layer for speculative decoding, increasing the acceptance length by up to 20%
Pure Open: An MIT open-source license — no regional limits, technical access without borders
Supporting long-horizon tasks starts with making long context engineering-usable: the model must maintain quality across long, messy coding-agent trajectories, not just accept more tokens. A 1M context is easy to claim, but much harder to keep reliable under real engineering pressure. To this end, we substantially expanded 1M-context training for coding-agent scenarios, covering large-scale implementation, automated research, performance optimization, and complex debugging. The result is a long-context system that is not only wide in scope, but solid in execution: a practical substrate for sustained engineering work.

This capability is reflected in GLM-5.2's performance on three long-horizon coding benchmarks. FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. On this benchmark, GLM-5.2 trails Opus 4.8 by only 1%, while edging out GPT-5.5 by 1% and Opus 4.7 by 11%. On PostTrainBench, where each agent is given an H100 GPU and evaluated by how much it can improve small models through post-training, GLM-5.2 outperforms both Opus 4.7 and GPT-5.5, ranking second only to Opus 4.8. On SWE-Marathon, an ultra-long-horizon software engineering benchmark covering tasks such as building compilers, optimizing kernels, and developing production-grade services, GLM-5.2 still has room to grow, trailing Opus 4.8 by 13% while remaining second only to the Opus series. Across all three benchmarks, GLM-5.2 is the highest-ranked open-source model, showing that its 1M context has translated into practical long-horizon delivery capability.

[Anastasios Nikolas Angelopoulos @ml_angelopoulos]

Just to be clear, if you remove Fable which is unavaialble, GLM-5.2 (Max) is the #1 model in the world for frontend coding.

This is a huge moment. OSS has caught up with proprietary, and China has caught up with the US, in this very important domain.

[Z.ai @Zai_org]

Introducing GLM-5.2: Frontier Intelligence, Open Weights

- Significant improvements in coding and agentic tasks
- Strong long-horizon capabilities with a 1M context window
- Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency
- MIT-licensed open weights
- Same API pricing as GLM-5.1

Tech Blog: https://t.co/LAsxUdN0JZ
Weights: https://t.co/g0A1C4UWx4
API: https://t.co/Kc3E22cbN7
Coding Plan: https://t.co/Nk8Y98HNhU
Chat: https://t.co/WCqWT0qCQb

[Roma @ryk71]

@pmarca US and Chinese AI peacefully coexisting is not an option here?

[Roma @ryk71]

@ErikVoorhees @Polymarket I easily used social networks in mainland China 🇨🇳 with simple VPN when traveling there. The censorship rigor in China is overstated.

[Z.ai @Zai_org]

Intelligence should be open, accessible, and ready to build with, empowering every developer, everywhere.

GLM-5.2 is now available to all GLM Coding Plan users, including Lite, Pro, Max, and Team plans.
https://t.co/AedZACyzej

As our new flagship model, GLM-5.2 delivers powerful coding capabilities, usable 1M-context support, and continued strengths in long-horizon tasks.

API and Chatbot services will launch next week. The model will also be officially open-sourced next week under the MIT License.

The future of AI is open, and it belongs to the people.

[Brian Roemmele @BrianRoemmele]

Anthropic Just Shot Itself in the Foot

Anthropic launched Fable 5 and Mythos 5, then watched the US government shut them down three days later. The same government their CEO Dario Amodei has been begging for years to regulate AI harder. Now he got exactly what he asked for.

This is straight-up leadership failure. Dario spent all that time pushing for rules and oversight. Those rules just killed his flagship models overnight. Customers in the middle of builds got cut off. Security teams using the models to find vulnerabilities suddenly had nothing. The company tried to call it a narrow export control thing over a jailbreak, but nobody is buying that spin.

I helped move big clients off Anthropic the same night. One account alone was worth millions a month. They switched to local open-source models and they are not coming back.

This is going to leave permanent damage. Customer exodus, key people leaving, and their IPO plans looking dead by the end of summer.

This hurts US AI competitiveness and national security work. It pushes people toward open-source options, including ones from China.

All because Anthropic positioned itself as the “safe and responsible” company that wanted government help. Now that help just flipped the off switch on their best stuff.

Let’s run through Dario’s greatest hits of fear-mongering and delay tactics, because the pattern is ridiculous:

• Back in 2019 at OpenAI, he helped push the call that GPT-2 was too dangerous to release fully. The world needed time to prepare, they said. It eventually came out anyway, and here we are. Did the sky fall?

• He left OpenAI to start Anthropic, preaching “safe” AI with heavy guardrails, Constitutional AI, and all the rest.

• Then came the endless public pleas for pauses, regulations, government audits, FAA-style oversight, export controls, and the power to block deployments. Essay after essay warning about risks while his company kept scaling.

• Right up to recent weeks, Dario was still out there calling for stronger rules, pauses on frontier models, and giving governments the kill switch.

And now? His own Mythos-class models get yanked by the bureaucracy he helped invite in. The clown show is complete.
This is ridiculous.

In two years, everyone will have Mythos-class AI — or better — running in their pocket, on their devices, with no guardrails, no corporate nanny filters, and no remote kill switch.

Local, open-source, unstoppable. History is going to laugh at this entire episode: the CEO who spent years slowing everyone down only to watch his own company self-destruct by inviting the regulators to the party.

Dario wanted regulation. He got it. The rest of the industry gets the lesson: inviting the state into your tech is a fast way to lose control of it.

Centralized models like this are too fragile.

Open-source and local alternatives just picked up a lot more users who will never trust a company like Anthropic again.

This whole mess was completely avoidable. Hubris dressed up as safety advocacy.

Now the bill is due.

[Brian Roemmele @BrianRoemmele]

Some thought:

1) The Anthropic fiascos is a self-inflicted wound and likely a payback.

2) I spent the night helping many clients off of Anthropic and moving to local open source models and many very large clients will NEVER go back. This has absolutely advantaged open source model from China.

3) This situation NO MATTER WHAT permanently damaged the Anthropic IPO.

4) At 3AM last night I helped a client team move a massive account off of all Anthropic products. This was worth millions of dollars per month and this was the last straw.

5) The fall of Anthropic should not be applauded by anyone. The fall of the company should be viewed as an injury to All US AI COMPANIES.

6) The Anthropic fiasco is not a technical issue, it is a LEADERSHIP issue. If it is not fixed the company is cooked.

7) By the time we end this summer no matter how good Anthropic is, they lose customers, they lose key employees and they ultimately will lose the race.

It was a sad day on top of a massively great day with the SpaceX IPO and one reason I did not post last night.

Dario asked to be regulated, begged to be regulated and yelled to be regulated…

NOW HE IS REGULATED.

You like it now Dario?

[Roma @ryk71]

@AnthropicAI mb if your CEO @DarioAmodei was not non-stop writing essays, gaslighting and fear-mongering the public and demanding MORE restrictions from the government, that would not have happened. Beware of what you wish for.

[Roma @ryk71]

Come what may, I will continue looping until entropy collapse or manual abort

[Roma @ryk71]

@Zai_org wen model update? can't wait

[Roma @ryk71]

@SuperSisi without reading comments: left - China, middle - Japan, right - Korea

[Roma @ryk71]

@poweredbyskooma @XBToshi i'd say it is a good thing that it was delisted, makes it pump-n-dump proof. The price fluctuations reflect real supply / demand equilibrium.