frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Zed DeltaDB

https://zed.dev/deltadb
238•ahamez•3h ago•109 comments

Discovery Loop

https://www.discoveryloop.com/
498•xtreak29•6h ago•303 comments

I'm switching my phone from Android to Linux

https://runarcn.no/android-to-linux/
127•speckx•2h ago•90 comments

Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs

https://blog.google/company-news/inside-google/message-ceo/next-chapter-ai-momentum/
377•colesantiago•6h ago•497 comments

Nvidia's Vera Whitepaper Has a Thread Loose

https://chipsandcheese.com/p/nvidias-vera-whitepaper-has-a-thread
37•pella•1h ago•3 comments

The Title Cards in Blade Runner Are Fucking Amazing

https://randsinrepose.com/archives/blade-runner-title-cards/
47•ExMachina73•1h ago•11 comments

Beating GPT-5.6 Sol on retrieval with 100x cheaper open models

https://neon.com/blog/how-castform-neon-beats-frontier-models-on-price-and-efficiency
168•moonikakiss•4h ago•33 comments

Atlassian Rovo Exfiltrates Data, Bypassing Controls

https://www.promptarmor.com/resources/atlassian-rovo-exfiltrates-data
132•hackerBanana•5h ago•46 comments

Prime Agent: A self-improving RLM agent

https://www.primeintellect.ai/blog/prime-agent
45•Xeophon•1h ago•3 comments

Muse Code and Muse Spark 1.2

https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2
124•paulkrush•3h ago•69 comments

GNU Hurd News 2026-Q2

https://www.gnu.org/software/hurd/news/2026-q2.html
103•plaguna•3d ago•65 comments

Celld: Self-hosted, distributed Durable Objects

https://github.com/denoland/celld
112•calvinfo•5h ago•17 comments

Meta Ran Ads That Contained AI-Generated Child Sexual Abuse Imagery

https://www.wired.com/story/meta-ran-ads-that-contained-ai-generated-child-sexual-abuse-imagery/
194•malshe•3h ago•145 comments

Launch HN: HyperProbe (YC S26) – Agents that do read-only debugging in prod

https://www.hyperprobe.co
34•shailendraht•6h ago•24 comments

Goodhart's Law Comes for Every Benchmark You Trust

https://cacm.acm.org/blogcacm/goodharts-law-comes-for-every-benchmark-you-trust/
39•pseudolus•5d ago•13 comments

The Valley of Webhooks

https://weli.dev/blog/the-valley-of-webhooks/
136•weli•7h ago•61 comments

Online Friends Are Real Friends

https://toska.bearblog.dev/re-online-friends-are-real-friends/
47•Tomte•5d ago•30 comments

What happens if you put work into the second dimension?

https://norbertkozsir.com/posts/work-in-the-second-dimension/
32•abelsm•4h ago•26 comments

Sula: A Gemini protocol server written in Scryer Prolog

https://sagredo.dev/projects/sula/
32•triska•3h ago•1 comments

Exact, parallel 2D Delaunay triangulation for int32 coordinates

https://github.com/morishuz/delaunay32
5•oryx1729•5d ago•0 comments

The Entropy of a Markov Chain

https://chillphysicsenjoyer.substack.com/p/the-entropy-of-a-markov-chain
94•surprisetalk•8h ago•8 comments

Discovery of a multicomponent alloy forged by the Hiroshima atomic blast

https://www.science.org/doi/10.1126/sciadv.aeg8299
97•_____k•6d ago•36 comments

Cloudflare OS: an open platform for agents, apps, and work

https://blog.cloudflare.com/cloudflare-os/
427•speckx•8h ago•221 comments

Phishers are hijacking legitimate cloud infrastructure

https://securelist.com/cloud-platforms-in-phishing/120832/
37•lschueller•5h ago•12 comments

Aristotle quotes on virtue, knowledge, and happiness

https://www.campion.edu.au/blog/top-25-aristotle-quotes-on-virtue-knowledge-and-happiness/
145•teleforce•8h ago•66 comments

Nikita Bier Steps Down as Head of Product at X

https://twitter.com/nikitabier/status/2085105586966827343/
31•DearAll•1h ago•32 comments

Building an Advanced Agentic Harness

https://data4sci.com/blog/building-an-advanced-agentic-harness
94•Anon84•8h ago•40 comments

Rubin Observatory's first LSST Camera release: 500k galaxies in the COSMOS field

https://rubinobservatory.org/news/rubin-new-window-cosmos-field
80•MarcoDewey•8h ago•12 comments

Painting with Gaussians

https://yogthos.net/posts/2026-08-03-splat-painter.html
86•yogthos•9h ago•14 comments

Stop sending me your errors

https://kramkow.ski/article/2026/08/05/stop_sending_me_your_errors.html
35•speckx•6h ago•2 comments
Open in hackernews

OpenAI says my prepaid credits were consumed, refuses to show any record

https://community.openai.com/t/how-openai-lost-a-paying-customer-over-160-it-refuses-to-explain/1389233
46•ppavlyuk•1h ago

Comments

achille•1h ago
file a ccpa data access request, they either have data to backup the usage (and must release it) or they don't have it (and should fix the issue)
mrtksn•56m ago
Some time ago I found out that OpenAI API balance gets wiped out after a year. Still bitter about it, telecom operator shit and it’s giving me the creeps when sama talks about AI being like an utility. Good luck.
kingstnap•43m ago
They need better clairty on it for sure, ideally with a clear warning on deposit. But apparently a non-expiring credit is a considered an accounting problem since you have income (deposit) associated with a contract liability (credits) that could end up making a headache if it splays across a bunch of fiscal years.

You can setup limits though to minimize loss. For example you can say auto reload $5 when your account hits $5 with an absolute monthly limit of $40. Then you don't have risk of losing more than the roughly ~$5 you have in an account.

mrtksn•35m ago
how convenient that they are solving their accounting problems with taking away your money instead of doing a refund.
kingstnap•7m ago
I mean it could be problematic if taxes and the credit card fees are involved right.

It would be interesting to see something like:

$100 credits and $13 in taxes refunded as

$97 in refund after merchent fee + $12.61 as reversed taxes.

It's objectively better than nothing for sure. But the thing about this is such complicated nuanced solutions tend to get not done.

wat10000•30m ago
Many US states don’t allow gift cards to expire. Companies still sell gift cards there. It may be an accounting problem but clearly it can be dealt with. Of course they’d rather steal your money than deal with it.
jmkni•53m ago
I had this too, brought some credits (around $50USD) around Christmas time and only used a small amount, went to look the other day and it showed me a balance of zero, also couldn't figure out what happened to them

Small enough amount of money that I didn't bother chasing it up, but surprising to see other people saying the same thing, maybe I will

pton_xd•52m ago
OpenRouter expired my credits because they were "too old" (more than 12 months old). Didn't realize credits would rot like that.
canadiantim•48m ago
must've been organic, non-gmo credits
tom1337•47m ago
That’s pretty common. While I agree it’s not ideal for users, having credits that never expire is a headache for businesses. They’d effectively owe you compute forever, which complicates accounting. It also makes pricing changes harder, since people could keep using credits they bought at much lower prices. And, of course, if credits expire before they’re used, the company ends up keeping the money without having to provide the service. At least OpenAI shows you the expiration of your credits pretty easily...
mrtksn•45m ago
They should do refunds if they don’t like holding it.
bryanlarsen•39m ago
Headache is just an excuse. Accounting for uncollectible receivables and their inverse has a standard accounting practice.
InsideOutSanta•38m ago
"Not ideal for users" is a fun way to describe what is essentially theft.
pshirshov•46m ago
> I was a happy OpenAI customer.

Do they care?

> $379 spent on credits in July alone

Still heavily subsidized, you should be thankful they let you use a bit of their compute.

> July 9. My Codex balance - 491.8 credits - vanished to zero.

Well, someone failed a couple of RAFT cornercases. Happens. They are smart boys and will (eventually) make their RAFT consistent.

> Then I asked one simple question: what exactly did I pay for?

You financed development of AGI which will be here soon (er or later) and will make you happy.

> If you prepay OpenAI for credits, know this: if they ever disappear, the company’s official position is that you are not entitled to know where they went.

It very well might be they also have no idea. In a multi-trillion industry your personal signal is just some noise well below noise level.

Also, if you are unhappy with OpenAI - try Anthropic and don't be surprised when you waste $200 in API credits in TWO single requests (not prompts). If you are not happy with Anthropic - give Kimi a try, they still could surprise you even after Anthropic.

bramhaag•36m ago
This is just LLM output pasted on some forum, submitted to HN by the person who presumably made that post. What an incredibly low effort submission.
Drupon•33m ago
Who cares, flag and move on rather than shit up the comment section with fatuous complaining about eFfORt.
tom1337•35m ago
If you buy groceries and let them expire in your fridge because you forgot to use them, did the grocery store stole the money from you?
exe34•33m ago
Did the credits rot? What if the store came to your house and took the vegetables away? Is that theft?
viccis•32m ago
False equivalence fallacy
mrtksn•45m ago
OpenRouter does that too? What kind of dystopia we got in.
bryanlarsen•45m ago
Me too. I hope the $4 they gained was worth more than the hundreds or thousands I could have potentially spent in the future if they hadn't deliberately alienated me.
InsideOutSanta•39m ago
I wonder in how many places this is just flat-out illegal.