Are the links wrong in that page? They seem to be posting to X.

Introducing attested.network: Proof of Payment for ATProtocol attested.network is an open spec for decentralized proof of payments on ATProtocol, built on what we learned making atprotofans.com. It formalizes the three-party attestation model and opens it up for any app to implement.

Writing about attested.network, an open spec for decentralized proof of payments on ATProtocol. It builds on what we learned making atprotofans.com. Draft is up and feedback is welcome.

Introducing Prefab 🎨

A generative UI framework for building MCP Apps, data dashboards, or whatever you (or your agent) want.

In Python.

(Really.)

100+ shadcn components. Real React. No JavaScript required. And built right into FastMCP 3.2.

prefab.prefect.io

This is very cool.

A Cryptography Engineer’s Perspective on Quantum Computing Timelines The risk that cryptographically-relevant quantum computers materialize within the next few years is now high enough to be dispositive, unfortunately.

Two papers came out last week that suggest classical asymmetric cryptography might indeed be broken by quantum computers in just a few years.

That means we need to ship post-quantum crypto now, with the tools we have: ML-KEM and ML-DSA. I didn't think PQ auth was so urgent until recently.

The problem I'm seeing with agentic driven development is that most models are trained on sample code or snippets and need very explicit skills to make sure the generated code is modular and sustainable. Sure, you can make changes that cross concerns very quickly now. But you shouldn't.

In that model, maybe success could be measured by quality and engagement. It could also be a nice way to filter out AI slop.

The network effect of being on the AT protocol is going to be real. I'm not sure if there's any value in measuring users on a specific PDS or relay? I could imagine a subscription membership model where posts by members can be set to limit interactions from other members.

Is network size per service a good measure of success if network effects aren't real in an interoperable system? Maybe better to measure quality, especially if subscription is the primary revenue stream and interaction on content is limited to paying members.

It's still not clear to me what the sustainable business model is on AT Protocol. When everything is interoperable there's always going to be an escape from advertising. Maybe that's the bigger question here - what is social media if the users aren't the product?

Have they talked about the business model somewhere? Doesn't monetization mean advertising?

If there's anything to learn from the Lean programming language in the age of agentic systems. Are types enough?

That's so true. But a disadvantage of using less OSS is that you don't get those ambient upgrades supplied by the community.

I'm convinced the same applies to smart TVs. Not battery life, just sluggish UI. There's no way this 4yo Samsung tv had one second lag on almost every interaction in the main menu.

Interesting, but note that the CES is US based. And center left would be classified as common sense, science based policymaking middle ground elsewhere. Like doing something about climate change, provide a social safety net.

Whatever shape the future economy has, AI should be a huge blow to enshitification
geohot.github.io//blog/jekyll...

JSSE is a new JS engine made by Claude over two months with four hours human supervision by @p.ocmatos.com

🦀 Choosing Rust aided the feedback loop
💯 Fully-compliant with test262(!)
🐌 Pure interpreter, unoptimized, slow
➰ Longest dev loop: 16 hours on Temporal
💸 ~$4k on Opus 4.6 API calls

A Visual Guide to Attention Variants in Modern LLMs From MHA and GQA to MLA, sparse attention, and hybrid architectures

What a great write up of LLM architectures. magazine.sebastianraschka.com/p/visual-att...

SQL, Typescript, and Agents How SQL’s Typescript moment will help agents just as much as it’ll help humans.

DBT Fusion as a structural type system for SQL to help coding agents. roundup.getdbt.com/p/sql-typesc...

I'm always puzzled when UI frameworks criticize React's state model design but then suggest a pattern that doesn't scale because it doesn't address asynchronous updates.

It's amazing how many data oriented teams are averse to SQL.

Investing in tests is a different case, but I think the goals are subtly different from a green field, language agnostic spec.

This may be rarely worth the tokens because waterfall development never worked. There's always going to be iteration and revision. Sure, you could remember the good prompts and run it again from scratch, but why waste the resources unless that's an explicit goal?

Judge orders Perplexity to stop AI agents from shopping on Amazon The AI browser fight heats up.

Hard to see how this is sustainable... Amazon winning in court to block AI agents using their site. The clock is ticking on placement advertising, and really all traditional web marketing. www.theverge.com/ai-artificia...

I think 'next token prediction' isn't really accurate, as it's also next sentence prediction and next thought prediction. Sure, those outouts are made of tokens - but that's too reductionist and hides these models' true capabilities.

GitHub - tc39/proposal-typedarray-concat: A proposal for optimizable TypedArray concatenation A proposal for optimizable TypedArray concatenation - tc39/proposal-typedarray-concat

It's going to be a busy week at TC-39 this week... I've got three proposals on the table I'll be trying to get advanced.

First is TypedArray concat github.com/tc39/proposa...

const u8_1 = new Uint8Array([1,2,3]);
const u8_2 = new Uint8Array([4,5,6]);
const u8_3 = Uint8Array.concat([u8_1, u8_2]);

Memory transfer time. Just goes to show there's no moat and no real lock in.

I've heard feedback on the @markpub.at lexicon! Thanks to @bmann.ca @blaine.bsky.social @thisismissem.social and others who have given really useful comments. I've published a new version for comment that is still very simple, but has more room for flexibility. Let me know what you think! markpub.at

Switch to Claude without starting over | Claude Transfer your preferences, projects, and context from other AI providers into Claude. Switch without losing what makes your AI useful.

No moat....
claude.com/import-memory

Right, basically Macbeth.