fp.

Frontier AI labs taking open-source and releasing it as a product

1•ainthusiast•59s ago•0 comments

LocalForge – Self-hosted LLM control plane with ML routing

https://github.com/al1-nasir/LocalForge

1•al1nasir•1m ago•0 comments

How we fixed prompt injection for all models on Fireworks

https://fireworks.ai/blog/safe-tokenization-preventing-prompt-injection-on-fireworks

1•smurda•2m ago•0 comments

Human drivers keep crashing into Waymos

https://www.understandingai.org/p/human-drivers-keep-crashing-into-454

2•speckx•3m ago•0 comments

Marathon training: Why hot baths might help you run faster

https://theconversation.com/marathon-training-why-hot-baths-might-help-you-run-faster-276796

2•PaulHoule•4m ago•0 comments

Cory Doctorow Admits Using a Slop-Machine to Produce His Drone of Propaganda

https://backlit.neocities.org/doctorow-admits-using-a-slop-machine-to-produce-his-propaganda

1•airhangerf15•4m ago•0 comments

Roo Code pivots to cloud-based agent, says IDEs aren't the future of coding

https://thenewstack.io/roo-code-cloud-ides-ai-coding/

1•Brajeshwar•4m ago•0 comments

Show IH: Knowledge Core

https://www.indiehackers.com/post/show-ih-knowledge-core-ecdf1646af

2•Ameya0•5m ago•0 comments

Show HN: Nimble.css minimal classless CSS library for great-looking default HTML

https://leftium.github.io/nimble.css/

2•Leftium•5m ago•0 comments

ChatGPT vs. a specialized medical AI on 5 clinical cases (verbatim outputs)

https://wizey.one/blog/2026/04/17/wizey-vs-chatgpt-5-clinical-cases-experiment/

2•wizeyone•6m ago•1 comments

Show HN: Reducing a 66-node dependency cycle to 13 in Scrapy

https://pvizgenerator.com/showcase/2026-04-scrapy-scc-refactor

2•pvizgenerator•6m ago•0 comments

High Intensity Functional Training: A Scoping Review (2025)

https://pmc.ncbi.nlm.nih.gov/articles/PMC12550923/

2•wslh•6m ago•0 comments

Browser using agents relying on expensive residential proxies?

2•maxalbarello•7m ago•0 comments

Anthropic: Using large language models to scale scalable oversight

https://www.anthropic.com/research/automated-alignment-researchers

1•theorchid•8m ago•0 comments

Show HN: Turning a Gaussian Splat into a Videogame

https://blog.playcanvas.com/turning-a-gaussian-splat-into-a-videogame/

2•yak32•9m ago•0 comments

'Staggering' number of people believe unproven claims about vaccines, raw milk

https://www.nature.com/articles/d41586-026-01285-2

3•voxadam•9m ago•0 comments

Teaching AI models to say "I'm not sure"

https://news.mit.edu/2026/teaching-ai-models-to-say-im-not-sure-0422

2•danroth27•10m ago•0 comments

The Sycophancy Problem: Why your AI is a Polite Liar (and how to fix it)

https://kampff.substack.com/p/the-sycophancy-problem-why-your-ai

2•Weatherill•10m ago•0 comments

A 95%-accurate AI agent fails 64% of the time on 20-step tasks

https://kenoticlabs.com/insights/ai-agent-failure

3•SamuelTanguturi•10m ago•0 comments

Paperback – generative UI, rendered on device – no pixel stream, no server GPU

https://bigballi.com/Paperback/

2•BigBalli•11m ago•0 comments

Mike Stonebraker Turing Award Winner: Postgres, Disagreeing with Google [video]

https://www.youtube.com/watch?v=YPObBOwIrHk

2•wrqvrwvq•11m ago•0 comments

Bitwarden CLI Compromised in Ongoing Checkmarx Supply Chain

https://socket.dev/blog/bitwarden-cli-compromised

4•tosh•11m ago•0 comments

Pentagon wants $54B for drones, more than most nations' military budgets

https://arstechnica.com/ai/2026/04/pentagon-wants-54b-for-drones-more-than-most-nations-military-...

3•voxadam•12m ago•0 comments

Show HN: Gova, a declarative GUI framework for Go

https://github.com/NV404/gova

3•vaibhavacharya_•12m ago•0 comments

Show HN: Claude Code skills for building LLM evals

https://github.com/latitude-dev/eval-skills

2•paulaq•13m ago•0 comments

OpenAI website has a full timeline of Elon Musk-related events

https://openai.com/elon-musk/

2•matthieu_bl•13m ago•0 comments

How to hit your Claude weekly limit so you can go outside and touch grass

https://jola.dev/posts/how-to-run-out-of-usage

2•shintoist•14m ago•2 comments

The Library Freedom Project

https://libraryfreedom.org/

2•lagniappe•15m ago•0 comments

Scripting macOS Image Playground to Create Game Assets for Free and Locally

https://martin.wojtczyk.de/2026/04/20/scripting-macos-image-playground-to-create-game-assets-for-...

2•wojtczyk•15m ago•0 comments

Medical data of half a million Britons listed for sale on Chinese website

https://news.sky.com/story/medical-data-of-half-a-million-britons-listed-for-sale-on-chinese-webs...

2•austinallegro•16m ago•1 comments

Frontier AI labs taking open-source and releasing it as a product

LocalForge – Self-hosted LLM control plane with ML routing

How we fixed prompt injection for all models on Fireworks

Human drivers keep crashing into Waymos

Marathon training: Why hot baths might help you run faster

Cory Doctorow Admits Using a Slop-Machine to Produce His Drone of Propaganda

Roo Code pivots to cloud-based agent, says IDEs aren't the future of coding

Show IH: Knowledge Core

Show HN: Nimble.css minimal classless CSS library for great-looking default HTML

ChatGPT vs. a specialized medical AI on 5 clinical cases (verbatim outputs)

Show HN: Reducing a 66-node dependency cycle to 13 in Scrapy

High Intensity Functional Training: A Scoping Review (2025)

Browser using agents relying on expensive residential proxies?

Anthropic: Using large language models to scale scalable oversight

Show HN: Turning a Gaussian Splat into a Videogame

'Staggering' number of people believe unproven claims about vaccines, raw milk

Teaching AI models to say "I'm not sure"

The Sycophancy Problem: Why your AI is a Polite Liar (and how to fix it)

A 95%-accurate AI agent fails 64% of the time on 20-step tasks

Paperback – generative UI, rendered on device – no pixel stream, no server GPU

Mike Stonebraker Turing Award Winner: Postgres, Disagreeing with Google [video]

Bitwarden CLI Compromised in Ongoing Checkmarx Supply Chain

Pentagon wants $54B for drones, more than most nations' military budgets

Show HN: Gova, a declarative GUI framework for Go

Show HN: Claude Code skills for building LLM evals

OpenAI website has a full timeline of Elon Musk-related events

How to hit your Claude weekly limit so you can go outside and touch grass

The Library Freedom Project

Scripting macOS Image Playground to Create Game Assets for Free and Locally

Medical data of half a million Britons listed for sale on Chinese website

The Atari 1200XL

Frontier AI labs taking open-source and releasing it as a product

LocalForge – Self-hosted LLM control plane with ML routing

How we fixed prompt injection for all models on Fireworks

Human drivers keep crashing into Waymos

Marathon training: Why hot baths might help you run faster

Cory Doctorow Admits Using a Slop-Machine to Produce His Drone of Propaganda

Roo Code pivots to cloud-based agent, says IDEs aren't the future of coding

Show IH: Knowledge Core

Show HN: Nimble.css minimal classless CSS library for great-looking default HTML

ChatGPT vs. a specialized medical AI on 5 clinical cases (verbatim outputs)

Show HN: Reducing a 66-node dependency cycle to 13 in Scrapy

High Intensity Functional Training: A Scoping Review (2025)

Browser using agents relying on expensive residential proxies?

Anthropic: Using large language models to scale scalable oversight

Show HN: Turning a Gaussian Splat into a Videogame

'Staggering' number of people believe unproven claims about vaccines, raw milk

Teaching AI models to say "I'm not sure"

The Sycophancy Problem: Why your AI is a Polite Liar (and how to fix it)

A 95%-accurate AI agent fails 64% of the time on 20-step tasks

Paperback – generative UI, rendered on device – no pixel stream, no server GPU

Mike Stonebraker Turing Award Winner: Postgres, Disagreeing with Google [video]

Bitwarden CLI Compromised in Ongoing Checkmarx Supply Chain

Pentagon wants $54B for drones, more than most nations' military budgets

Show HN: Gova, a declarative GUI framework for Go

Show HN: Claude Code skills for building LLM evals

OpenAI website has a full timeline of Elon Musk-related events

How to hit your Claude weekly limit so you can go outside and touch grass

The Library Freedom Project

Scripting macOS Image Playground to Create Game Assets for Free and Locally

Medical data of half a million Britons listed for sale on Chinese website