These 20 Awesome API Clients Will Change How You Work with APIs
Postman has become a staple in API development, but it’s far from the only option. As the API...
news/faster-builds
A delightful language with friendly error messages, great performance, small assets, and no runtime exceptions.
Create MCP servers with the Postman API Network | Postman Docs
With Postman’s MCP Generator, you can create a Model Context Protocol (MCP) server with public APIs from the Postman API Network. Then, you can use Postman to experiment, test, and evaluate your MCP server and an AI model of your choice.
First Languages AI Reality | Mila
The FLAIR initiative aims to develop a method for the rapid creation of customized ASR models for Indigenous languages.
Co-Writing with Opinionated Language Models Affects Users’ Views
As part of the Digital Library's transition to Open Access, new features for researchers are available in the Premium Edition. Click here to learn more.
Co-Writing with Opinionated Language Models Affects Users’ Views
As part of the Digital Library's transition to Open Access, new features for researchers are available in the Premium Edition. Click here to learn more.
Reddit API controversy
In April 2023, the discussion and news aggregation website Reddit announced its intentions to charge for its application programming interface (API), a feature which had been free since 2008, causing a dispute. The move forced multiple third-party applications to shut down and threatened accessibility applications and moderation tools.
ICML Poster Learning syntax without semantics: Disentangled tiny language models
Language models acquire syntax and world knowledge together, entangling the two in ways that limit efficiency and controllability. We show that syntax can be learned while suppressing semantic plausibility and world‑knowledge cues, yielding more efficient and controllable models. We train tiny LMs on grammatical nonsense — syntactically well-formed text with semantic content ablated via constrained relexicalization (SAMBAL). Models trained on this data perform comparably to standard pretraining on syntactic benchmarks (BLiMP, SyntaxGym) while scoring at chance on world knowledge probes (EWoK). On targeted grammar-plausibility conflict probes, content-neutral models prefer grammaticality where standard models prefer plausibility, and their representations show more syntactic vs lexical alignment. On efficiency, disentanglement yields substantial sample and parameter gains: in low‑resource regimes, a 5M‑parameter model matches a 30M‑parameter baseline at the same data budget. On controllability, content-neutral models adapt rapidly to a new domain with minimal exposure, suggesting the feasibility of modular post‑hoc knowledge specialization.
Function calling | OpenAI API
Learn how function calling enables large language models to connect to external data and systems.
to add to this, there’s a few product-y reasons why I would really like a createPost API 1. we could run a server-side language detection model that actually works, so we wouldn’t need the language select in the client (it sucks) 2. you wouldn’t need to specify a createdAt time anymore (cont.)
P(aul) Frazee
Nerding out about the question of whether the client or the server should be the one who speaks to the PDS
Nov 13, 2025 at 10:11 PM
This is partly why I think atproto needs a more convincing vision about being a post-app protocol. A future where atproto is a bunch of apps and each app governs their own lexicon diminishes a lot of its potential in the long-term.
Laurens
yeah i think i might actually, theres a very interesting story here that the main thing that matters is a focal point around which consensus over a lexicon can get formed
i wonder how much of this can be prototyped in userland with no new ui eg just a thing that creates “collection” accounts operated by bots. curators can post stuff there by dm’ing the bot. as a reader, you follow the account, so naturally it plays into your Following and For You etc
dan
“semble for bluesky inside bluesky app” is a good way to articulate exactly what i’m missing in the product. custom feeds are like half of that but they’re a readonly interface
i wonder how much of this can be prototyped in userland with no new ui eg just a thing that creates “collection” accounts operated by bots. curators can post stuff there by dm’ing the bot. as a reader, you follow the account, so naturally it plays into your Following and For You etc
dan
“semble for bluesky inside bluesky app” is a good way to articulate exactly what i’m missing in the product. custom feeds are like half of that but they’re a readonly interface
What if a PDS became more than a repository for posts? We propose using AT Protocol as a creator-controlled publication layer for declaration metadata describing digital works, making rights, provenance, and other trusted metadata independently verifiable and easier to discover.
From Personal Data Server to Personal Metadata Server
liccium.leaflet.pubThe computer now has a natural language interface and a replicator for software You think I’m going to do anything other than ensure this technology gets into as many people’s hands as possible? The fight isn’t whether we use this. It’s whether we can self host these models
posts inspector
I am basically radicalized about this. We fight now for personal computing and personal agency or we lose another decade at best
