frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Genie 4

https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/
658•acossta•7h ago•402 comments

Ghostty 1.4

https://ghostty.org/docs/install/release-notes/1-4-0
657•deaux•6h ago•204 comments

Tailscale SSH: Investigation and remediation update

https://tailscale.com/ssh-investigation-update
154•figmert•3h ago•91 comments

Survival LLM: AI assistant and emergency library that runs on your own computer

https://www.brangus.io/survival-llm
135•Luc•3h ago•86 comments

Early progress on accelerating drug development with Claude

https://www.anthropic.com/research/early-progress-drug-development
77•dataking•2h ago•69 comments

Raspberry Pi Zero 3W and Pi 4 Model A spotted in device trees

https://bret.dk/raspberry-pi-zero-3-w-pi-4-model-a/
67•sthlmb•2h ago•19 comments

SS United States to be sunk as largest artificial reef

https://divemagazine.com/scuba-diving-news/ss-united-states-to-be-sunk-as-worlds-largest-artifici...
52•geox•1d ago•37 comments

New 14‑inch MacBook Pro with M6

https://www.apple.com/in/newsroom/2026/10/apple-introduces-the-new-14-inch-macbook-pro-with-m6/
253•beex•7h ago•305 comments

Four Months of Zig

https://gracefulliberty.com/articles/four-months-of-zig/
52•ibobev•3h ago•25 comments

An examination of lost firefly species of the US and Canada

https://journaloflostspecies.org/index.php/jls/article/view/37/33
69•bryanrasmussen•3d ago•9 comments

Reviewers rarely talk about reliability. Owners bring it up twice as often

https://www.bettascore.ai/research/reliability-reviewers-vs-owners
12•tsxng•1h ago•15 comments

Python's experimental JIT to be removed

https://lwn.net/SubscriberLink/1102102/0c00e4d550f25f8c/
427•f311a•12h ago•220 comments

Classification of Finite Simple Groups Formalized in Lean

https://cfsg.axiommath.ai/
155•pfdietz•7h ago•41 comments

New iMac with M6

https://www.apple.com/imac/
295•notfried•10h ago•353 comments

Wikipedia's Cleanup of AI-Generated Content

https://spectrum.ieee.org/wikipedia-wikiproject-ai-writing-cleanup
429•rbanffy•12h ago•198 comments

Radio-tagged Monarch butterflies

https://news.vt.edu/articles/2026/10/cmi-monarchs.html
14•rdmuser•5h ago•5 comments

Introducing LFM3: Every Modality for Specialized Intelligence

https://www.liquid.ai/blog/lfm3
105•Philpax•6h ago•39 comments

Hex Raises $100M

https://hex.tech/blog/hex-new-era/
28•tomwphillips•3h ago•12 comments

US Government is hiding a secret court ruling about how it spies on Americans

https://www.wired.com/story/the-us-government-is-hiding-a-secret-court-ruling-about-how-it-spies-...
728•helterskelter2•9h ago•263 comments

Nvidia and Micron Make Up 34% of America's Earnings Boom

https://www.wsj.com/finance/stocks/the-earnings-boom-is-heavily-reliant-on-just-a-handful-of-stoc...
30•JumpCrisscross•2d ago•3 comments

Three TLS Bugs in GoAhead

https://foxio.io/blog/choose-your-backend-three-tls-bugs-in-goahead
50•owenshen24•5h ago•8 comments

Land Value Capture in Practice: Ethiopia's Urban Land Lease System

https://blog.landeconomics.org/p/land-value-capture-in-practice-reflections
32•surprisetalk•1d ago•3 comments

Why Micron Is Suddenly Ditching 4F² DRAM as Rivals Double Down

https://hothardware.com/news/why-micron-is-suddenly-ditching-4f-dram
128•rbanffy•1d ago•20 comments

One year of Dioxus fullstack in production: the good, the painful

https://oxidt.com/blog/one-year-of-dioxus-fullstack-in-production-the-good-the-painful
17•hauju•2h ago•10 comments

Discord will soon let anyone find public server messages through Google

https://discord.com/blog/making-content-from-discoverable-servers-easier-to-discover
199•DeepLogin•4h ago•114 comments

Anthropic trained Claude on illegally obtained Stanford professors' books

https://stanforddaily.com/2026/10/27/professors-books-pirated-to-train-claude/
422•car•8h ago•236 comments

Amazon Is Gutting Its AI Division After Sustained Failure

https://futurism.com/artificial-intelligence/amazon-ai-division-failure-nova
212•dpflan•7h ago•113 comments

Here be monsters: What do you still need to know when agents do the work?

https://blog.adrq.dev/here-be-monsters-what-do-you-still-need-to-know-when-agents-do-the-work
143•adrq•10h ago•73 comments

Rosettes and Other Hidden Traces of Hamburg's Tram

https://alsterrunde.com/rosettes-and-other-hidden-traces-of-hamburgs-tram/
29•mertbio•1d ago•1 comments

PHP2Shell, LiteLLM, GitLab: Find What You Need to Patch

https://cloudsleuth.io/blog/php2shell-litellm-gitlab-find-what-you-run/
54•bigfish2026•7h ago•9 comments
Open in hackernews

Reviewers rarely talk about reliability. Owners bring it up twice as often

https://www.bettascore.ai/research/reliability-reviewers-vs-owners
12•tsxng•1h ago

Comments

jdw64•49m ago
Because reviewers don’t use things for long, while owners usually do. So reviewers generally tend to fixate on new features by default.
lstodd•46m ago
Please don't normalize incompetence.
jdw64•36m ago
Rather than taking incompetence for granted, it’s because—as the article also says—motivations differ, and it’s difficult to judge factors like warranty handling or updates. And fundamentally, “reliability” is broad. That’s even more the case for combined hardware-software products. Ultimately, Reddit is basically where people with complaints go, and reviewers themselves have bias too, but consumer complaints also largely come from people whose products broke, so it’s difficult to judge.

But ultimately, for reviewers, what matters is that they “make a living through reviews,” so they have to say things that fit the fanboy tier they’re targeting.

The problem is that fanboys are generally people who buy more devices to begin with, so they also judge reliability based on the existing product lineup, and reviewers are largely the same. As a result, they can’t help but judge mainly by new features.

To be precise, “reliability” requires predicting the future, and that’s not easy. How would you know if your computer might break tomorrow? There are serious hinge issues with the iPhone Duo right now—how would people buying it have predicted that?

Prediction is basically break-even even when you’re right, and easy to criticize when you’re wrong, so it’s a hard attribute for reviewers to review.

florbo•33m ago
I understand your point but I can't expect a single sample review for a new product to contain anything useful in regards to reliability.
anigbrowl•32m ago
Ridiculous argument. Reviewers' task is to evaluate new products to provide information to consumers in a timely fashion. There's a niche for 'long term reviews' but it's tiny by comparison, and mainly sought out by nerds who are unusually interested in that product category.

Most reviewers simply can't evaluate products over a sufficiently long term to make convincing judgements about reliability. Where the latter is important, it's better to consult a source like consumer Reports that specializes in that sort of longitudinal assessment at the price of narrow product focus and shallower review detail.

tsxng•32m ago
Lol. Fair, though most reviewers only get a unit for a few weeks.
jasonlotito•5m ago
Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something.

https://news.ycombinator.com/newsguidelines.html

arealaccount•36m ago
Is this like Amazon where the reviewers are basically on a “get free stuff” list, and they are no longer selected once they leave a negative review?’
tsxng•24m ago
Probably part of it. A lot of reviewers get their units for free and it's easier to be happy with something you didn't pay for. We can't tell which reviews were sponsored, so we can't measure that directly.
arealaccount•15m ago
“Sponsored” is an interesting word to use
CPUHall•18m ago
Yeah well not many reviewers actually use them long enough before putting up a review - fair play, they review what they have on hand. Reviewers are mostly working a job too, the money boat is during a release of a product, not a year after they've well tinkered with it - assuming they even have the time to with multiple same-category products released constantly.

Anyway, all i'm saying is its not terrible. People who watch reviews on release should know there is no reliability to talk about unless the reviewer spends the effort to 'accelerate' these - e.g. Fold testing foldable phones and counting them.

Take a look at my country's (Singapore) arguably #1 e-commerce platform and you'll notice something even more mindblowing - platform-generated, ai-generated reviews. And no, its not the sellers, most of them have even voiced their frustrations that they dont even have an option to block these 'ai-generated reviews' on their products. Yes, there are ai-generated humans wearing ear0hones clipping through ears, on the complete opposite directions and telling you how awesome the product is. Truly dystopian.

cwillu•18m ago
> .bs-consent__sheet { … }

Couldn't have said it better myself

jedbrooke•8m ago
tangentially related: I’ve been apartment hunting recently and it reminds me of a common occurrence for apartment reviews:

5 star reviews are all: “I did the tour and the agent was friendly/place looked nice”

meanwhile the 1 star reviews: “I actually live here and this place is terrible.”

no doubt they ask you to review 5 stars on the tour…