Feed

That era of building a viable SaaS business in a few months is gone. I mean, it technically still exists today but only in the arbitrage sense that the rest of the world hasn’t yet caught on to how quickly and easily software can be built. It’ll be gone soon, I promise.

In atproto, every app is a projection of the whole Atmosphere, just like Feedly and Google Reader are projections of the entire Blogosphere. You mostly “decentralize” by swapping your hosting, and/or by making and trying new apps. Running many full copies of the Bluesky database server is possible, but it’s not any more useful than running many copies of Google Reader.

Here’s a partial list of scam-relevant capabilities that LLMs have that would previously require significant skilled human effort per target:

Researching a mark to find out the best way to go after them.
Personally tailoring all communication with a mark in mind, dynamically adjusting based on how they respond to various approaches.
Cloning the voice of a person the mark trusts, like a relative.
Plausible, real-time deepfaking of a video call.
Building a plausible-looking corroborating fake web presence 6.
Realtime monitoring of compromised resources, and dynamically building up the scam based on this monitoring.
Better triage and discovery of marks.
Avoiding signature-detection based spam filters (shown by Heiding et al).
Scanning for and chaining known exploits in unpatched deployed software. Mass scanning isn’t new, but cheaply building tooling that can keep tabs on the latest CVEs and learn new tricks is7.

These are capabilities that exist today, and they’ll only improve from here. We should look at these skills as a floor, not a ceiling.

The user opens the URL on whatever device they want, not necessarily the box running the CLI. They log in, see the requested scopes and the client name, confirm the short code matches what the CLI printed, and approve.

Gemma-4-12b-qat just came out but I’ve already also really been impressed with its performance relative to its size. The model architecture itself is really interesting and proposes a bunch of interesting questions like, “if we are constrained by performance and price, what architectural tradeoffs do we need to make?” a question that so far has not really been asked in the mad token gold rush.

Will anything survive in roughly its current form?

Probably. Experience that isn’t solely information: comedy, entertainment, storytelling, fiction, etc. You don’t ask an AI to summarize a stand-up special, and a synopsis of a great novel is not a great novel. Voice, taste, and personality may end up being the only durable moats. But “give me the 5 steps to X”? That’s a tough business that’s about to get a lot tougher.

“Casey can use the system to communicate his own thoughts, not only while we're there in a controlled environment, but whenever he wants. Sometimes, he would do that over 12 straight hours,” said the study’s lead author, Nicholas Card. Card is a postdoctoral scholar in the UC Davis Department of Neurological Surgery. “The system worked well, was reliable and stable, and delivered consistent results. This is one of the strongest demonstrations that BCIs can be practical and useful.”

The main conclusion was that the MTP draft model is worth using. On this machine it took Gemma 4 from 58.2 tokens/second to 72.2 tokens/second, while keeping the setup simple enough to run as a local OpenAI-compatible server.

Standard.Site aims to do something similar for any platform powered by the AT Protocol. The Standard.Site schema provides you with a set of lexicons you use to describe your website and the individual pages and posts on it in a way where they can be parsed by a service that runs on the AT Procolo — an “App View,” in the language of their standard — which would allow users to subscribe, via that service, to your domain. The most obvious use case is allowing users on a platform like Bluesky to subscribe to an RSS-like feed of, say, blog posts, the way they would follow any other account on the service.

A compiler does not feel humiliated when I swear at it, a car does not suffer when I call it a shitbox and a power drill is not oppressed by being handled roughly. An LLM is more complicated than those things, and the interactions you can have with them can be truly uncanny, but a moral status does not appear just because the machine can emit text in the first person.