frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Hoot: Scheme on WebAssembly

https://www.spritely.institute/hoot/
2•AlexeyBrin•1m ago•0 comments

What the longevity experts don't tell you

https://machielreyneke.com/blog/longevity-lessons/
1•machielrey•2m ago•0 comments

Monzo wrongly denied refunds to fraud and scam victims

https://www.theguardian.com/money/2026/feb/07/monzo-natwest-hsbc-refunds-fraud-scam-fos-ombudsman
2•tablets•7m ago•0 comments

They were drawn to Korea with dreams of K-pop stardom – but then let down

https://www.bbc.com/news/articles/cvgnq9rwyqno
2•breve•9m ago•0 comments

Show HN: AI-Powered Merchant Intelligence

https://nodee.co
1•jjkirsch•12m ago•0 comments

Bash parallel tasks and error handling

https://github.com/themattrix/bash-concurrent
2•pastage•12m ago•0 comments

Let's compile Quake like it's 1997

https://fabiensanglard.net/compile_like_1997/index.html
1•billiob•13m ago•0 comments

Reverse Engineering Medium.com's Editor: How Copy, Paste, and Images Work

https://app.writtte.com/read/gP0H6W5
2•birdculture•18m ago•0 comments

Go 1.22, SQLite, and Next.js: The "Boring" Back End

https://mohammedeabdelaziz.github.io/articles/go-next-pt-2
1•mohammede•24m ago•0 comments

Laibach the Whistleblowers [video]

https://www.youtube.com/watch?v=c6Mx2mxpaCY
1•KnuthIsGod•25m ago•1 comments

Slop News - HN front page right now hallucinated as 100% AI SLOP

https://slop-news.pages.dev/slop-news
1•keepamovin•30m ago•1 comments

Economists vs. Technologists on AI

https://ideasindevelopment.substack.com/p/economists-vs-technologists-on-ai
1•econlmics•32m ago•0 comments

Life at the Edge

https://asadk.com/p/edge
3•tosh•38m ago•0 comments

RISC-V Vector Primer

https://github.com/simplex-micro/riscv-vector-primer/blob/main/index.md
4•oxxoxoxooo•41m ago•1 comments

Show HN: Invoxo – Invoicing with automatic EU VAT for cross-border services

2•InvoxoEU•42m ago•0 comments

A Tale of Two Standards, POSIX and Win32 (2005)

https://www.samba.org/samba/news/articles/low_point/tale_two_stds_os2.html
3•goranmoomin•45m ago•0 comments

Ask HN: Is the Downfall of SaaS Started?

3•throwaw12•46m ago•0 comments

Flirt: The Native Backend

https://blog.buenzli.dev/flirt-native-backend/
2•senekor•48m ago•0 comments

OpenAI's Latest Platform Targets Enterprise Customers

https://aibusiness.com/agentic-ai/openai-s-latest-platform-targets-enterprise-customers
1•myk-e•51m ago•0 comments

Goldman Sachs taps Anthropic's Claude to automate accounting, compliance roles

https://www.cnbc.com/2026/02/06/anthropic-goldman-sachs-ai-model-accounting.html
3•myk-e•53m ago•5 comments

Ai.com bought by Crypto.com founder for $70M in biggest-ever website name deal

https://www.ft.com/content/83488628-8dfd-4060-a7b0-71b1bb012785
1•1vuio0pswjnm7•54m ago•1 comments

Big Tech's AI Push Is Costing More Than the Moon Landing

https://www.wsj.com/tech/ai/ai-spending-tech-companies-compared-02b90046
4•1vuio0pswjnm7•56m ago•0 comments

The AI boom is causing shortages everywhere else

https://www.washingtonpost.com/technology/2026/02/07/ai-spending-economy-shortages/
2•1vuio0pswjnm7•58m ago•0 comments

Suno, AI Music, and the Bad Future [video]

https://www.youtube.com/watch?v=U8dcFhF0Dlk
1•askl•1h ago•2 comments

Ask HN: How are researchers using AlphaFold in 2026?

1•jocho12•1h ago•0 comments

Running the "Reflections on Trusting Trust" Compiler

https://spawn-queue.acm.org/doi/10.1145/3786614
1•devooops•1h ago•0 comments

Watermark API – $0.01/image, 10x cheaper than Cloudinary

https://api-production-caa8.up.railway.app/docs
2•lembergs•1h ago•1 comments

Now send your marketing campaigns directly from ChatGPT

https://www.mail-o-mail.com/
1•avallark•1h ago•1 comments

Queueing Theory v2: DORA metrics, queue-of-queues, chi-alpha-beta-sigma notation

https://github.com/joelparkerhenderson/queueing-theory
1•jph•1h ago•0 comments

Show HN: Hibana – choreography-first protocol safety for Rust

https://hibanaworks.dev/
5•o8vm•1h ago•1 comments
Open in hackernews

The Continual Learning Problem

https://jessylin.com/2025/10/20/continual-learning/
68•kiyanwang•3mo ago

Comments

mynti•3mo ago
Super interesting blogpost. I just wonder how this is actually different to LORA, since LORA also adds some parameters and freezes the rest of the model. This seems like a sparse, memory efficient LORA with a couple of extra steps, since it uses attention again to make the sparsity work. All while making it a lot more effective compared to LORA (performance drop of only 11% compared to 71%).
sva_•3mo ago
> LORA

I think you meant LoRA (not to be confused with LoRa)

alyxya•3mo ago
I think the solution to continual learning is as simple as using context distillation. We know that models are good at in-context learning, so we just want an efficient way to distill context into the weights. I suspect context rot may come from how the softmax in attention gets diluted with a longer context, so this wouldn't be an issue with context distillation.
killerstorm•3mo ago
Perhaps it can work through multiple stages: ICL -> prompt/context optimization (*) -> prefix tuning / KV distillation -> context distillation.

*: it is possible to measure how much part of a prompt helps with a task e.g. measuring change in entropy

imtringued•3mo ago
The problem with continual learning is that stochastic gradient descent is already an online algorithm applied incrementally on a shuffled dataset. If you add new data, you can't train on just the new data, because you will be running what amounts to a completely different training sequence. Further training requires the old data and the new data to be shuffled together.

With reinforcement learning, specifically actor critic, the actor is not training against a dataset. It's training against the critic. The critic is supposed to approximate the value function, which contains the current cost for a given action and the predicted future cost, assuming that you choose the optimal action at every step, including its impact on future actions. If you have a simple supervised cost function, what happens is that the critic acts as an averaging of loss functions. You could say that the critic is a compressed copy of the training data. When you train the actor, you're essentially taking not only the new data, but also the old data into account.

So, in a way, catastrophic forgetting is sort of solved, but not really. If you add new data, you run into the problem that your critic will slowly drift to the new data distribution. This means the problem wasn't solved, but you certainly managed to delay it. Delaying the problem is good though. What if you can delay it even more? What if you can delay it forever?

Here is my stupid and simple unproven idea: Nest the reinforcement learning algorithm. Each critic will add one more level of delay, thereby acting as a low pass filter on the supervised reward function. Since you have two critics now, you can essentially implement a hybrid pre-training + continual learning architecture. The most interesting aspect here is that you can continue training the inner-most critic without changing the outer critic, which now acts as a learned loss function.