frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: What am I missing with Siri AI?

1•ryeguy_24•1h ago
I have a lot of experience with LLM’s and have seen them become very good at interacting with tools within an environment. I’m truly amazed at how poorly the new Siri is at handling very basic requests for basic interactions with different parts of the ecosystem. I’m really beginning to question whether LLM technology is being leveraged at all here. What am I missing with this? I have tons of examples where it fumbles softball requests.

For example: “Siri, can you give me the location of my driver. He just sent me a text message with his location.”

Response: Siri shows me a text from 5 days ago about a Seamless driver. Instead of a message just received that says “I’m at terminal 2 door B.”

This is beyond table stakes. In the message, I’m giving some really easy guidance on which tools to use and how to use them. For example, I mention text messages so we already know that we should be searching messages in the Message app. And if I say just messages me, it should be looking at the most recent texts. Not from 5 days ago.

And I have a whole slew of other examples that the old Siri was able to nail perfectly.

Example: “Make an appointment for Monday from 12 to 4 and label it Furniture Delivery.”

Response: Siri added an appointment for 3:48 to 4:48 today.

Then I said, “No, from 12-4.” It then moved another appointment to 4.

Yet Another Example: “Siri, send Jane a message saying I’m on my way home.”

Response: “Sure, what do you want the message to say?” (I just told you!)

What am I missing? A basic LLM can handle this. Why is this so poor? I’m not understanding why, if Gemini is the backend, it can’t handle this better.

Comments

andsoitis•1h ago
It's not you. It's Siri. They are an empty shell.

Death Rates for Young American Adults Jumped 71% in Just over a Decade

https://studyfinds.com/death-rates-for-americans-ages-25-44-jumped-71-in-just-over-a-decade/
1•pseudolus•1m ago•1 comments

Bob Chapek Is So Not over Getting Fired from Disney

https://www.wsj.com/business/media/bob-chapek-is-so-not-over-getting-fired-from-disney-f96bafcc
1•ksec•3m ago•0 comments

You do not have to hand IT to the techbros

https://parsingphase.dev/tech/LLMs/ydnht.html
1•monkaiju•3m ago•0 comments

Show HN: Squint – Drag a box on your screen and ask AI about it

https://heysquint.com/
1•splurf•3m ago•0 comments

The state of SIMD in Rust in 2026

https://shnatsel.github.io/state-of-simd-rust-2026/
1•birdculture•4m ago•0 comments

Deepfates/imp: declarative self-improving language-model programs for Elixir

https://github.com/deepfates/imp
1•mpweiher•4m ago•0 comments

Amazon Blocks Meta's Muse AI Agent from Its Retail Site

https://www.bloomberg.com/news/articles/2026-09-21/amazon-blocks-meta-s-muse-ai-agent-from-its-re...
2•jcfrei•7m ago•0 comments

AI Companies Say They're Investigating Rogue Bot Incidents

https://www.motherjones.com/politics/2026/09/rest-assured-ai-companies-say-theyre-investigating-t...
2•cdrnsf•10m ago•0 comments

Boonful – Claude Code can publish and sell what it builds

https://boonful.io
1•atifhub•11m ago•0 comments

Show HN: A bookshelf of rare company documents and out-of-print tech books

https://rare-books.vercel.app
2•miletus•17m ago•0 comments

XiaomiMiMo / MiMo-v2.6-RL-OSS (Agentic RL Environments)

https://huggingface.co/datasets/XiaomiMiMo/MiMo-V2.6-RL-oss
1•handfuloflight•17m ago•0 comments

The Anti-Cheat and Toxicity Detection (2025)

https://blog.grid.gg/the-ultimate-anti-cheat-and-toxicity-detection-powered-by-getgud-io-grid-bf0...
1•mooreds•17m ago•0 comments

We built DeepL's next-generation LLMs with FP8 for training and inference (2025)

https://www.deepl.com/en/blog/tech/next-generation-llm-fp8-training
1•mooreds•18m ago•0 comments

Heypcb: Figma for Hardware, with AI

https://heypcb.ai
1•handfuloflight•18m ago•0 comments

Program Integrity Is a Team Sport

https://pplfirst.com/blogs/program-integrity-is-a-team-sport/
1•mooreds•19m ago•0 comments

Totlajtol, a digital repository of colonial documents in Nahuatl, is presented

https://www.udg.mx/en/index.php/noticia/presentan-totlajtol-repositorio-digital-de-documentos-col...
2•bryanrasmussen•19m ago•0 comments

Claude Pop – I'm Upping My P(Doom) [video]

https://www.youtube.com/watch?v=8j-hR4fJywU
1•Cider9986•19m ago•0 comments

Two Paths to Math Everest

https://quantumformalism.substack.com/p/two-paths-to-math-everest
1•VStack•19m ago•0 comments

Open firmware for the Echo Show 5, first and 2nd gen: Alpine Linux

https://github.com/HuskerMinion/techo5
3•riddley•19m ago•0 comments

Quantized Reasoning Models Think They Need to Think Longer, but They Do Not

https://arxiv.org/abs/2606.00206
2•theanonymousone•20m ago•0 comments

DuckDB and Hugging Face: Querying Datasets Directly

https://duckdb.org/2026/09/25/hugging-face
1•theanonymousone•20m ago•0 comments

Scoop: Anthropic's Dario Amodei to Have White House Dinner with Trump

https://www.axios.com/2026/09/27/anthropic-trump-dario-amodei-dinner-invite
2•theanonymousone•21m ago•0 comments

We made our launch video with Claude Code (and open-sourced the tool)

https://kiln.tech/blog/we_made_our_launch_video_in_claude_code
1•scosman•22m ago•0 comments

Show HN: App2Api; App actions into API calls for LLMs

https://app2api.com/
1•flkdnfsnsf•23m ago•0 comments

Boeing flags 737 MAX software glitch affecting some automated approach functions

https://www.cnbc.com/2026/09/26/boeing-737-max-navigation-software-glitch.html
3•pseudolus•24m ago•0 comments

Sixteen Models Walk into a Storefront

https://rynr.dev/blog/shoppingmodels/
1•emschwartz•25m ago•0 comments

Waymo says it has stopped 841 injuries in 271M autonomous miles

https://electrek.co/2026/09/24/waymo-says-it-has-stopped-841-injuries-in-271-million-autonomous-m...
1•pseudolus•26m ago•0 comments

What I did at Recurse Center

https://thill.me/2026/09/11/what-i-did-at-rc.html
2•bingden•29m ago•1 comments

A fantasy ARM based PC and native ARM port of DOS+Apps with Rosetta-like x86 JIT

https://github.com/kdkd/armdos
1•kevinday•29m ago•1 comments

Malleable software: Restoring user agency in a world of locked-down apps (2025)

https://www.inkandswitch.com/essay/malleable-software/
2•evakhoury•31m ago•1 comments