Too slow: latency and long waits

In contrast, someone who is happy to consume modern technology accept AI for arbitrary reasons seems like someone who is making interacting with society unnecessarily complicated and difficult. >Or that his buggy is too slow for the roads (not highways), etc.

nater5000 · Oct 28, 2026 · comment in “Kentucky man becomes the first humanist granted religious exemption from AI”

To support high performance git in FUSE, our writes hit remote NVMe disk(s) and asynchronously sync to S3 (you can replicate on NVMes or just do failure recovery). Previously, we used remote NVMe disks as a write-through cache, but that doesn't work with git, S3 latency is too high.

jamesblonde · Oct 11, 2026 · comment in “How big is a Git commit?”

I also find that it's easy to spend like that, but also easy not to with little impact on productivity. At the current moment I'm stuck with rider + copilot (not ideal), but e.g. using GPT 6.1 luna is really, really cheap, and lots of tasks are quickly and decently dealt with even at lower reasoning levels, (added bonus of having low latency).

emn13 · Oct 9, 2026 · comment in “Why isn't the industry freaking out about DeepSeek 4.1 Flash?”

It isn't just driven by the AI companies, either. I've already seen it internally, with developers who need diagnostic tools creating Claude skills, and in the excitement of adopting AI, barely noticing that they're incredibly slow, expensive, unreliable, and even slow to create compared to the script or diagnostic endpoint they would created in the past.

dkarl · Oct 3, 2026 · comment in “Every SaaS business will become a harness around a model”

I have been doing it since the first chain of thought open models became available. perhaps there may be a way to get a Jev-type model to think for a very specific number of steps to gain control over its latency, if so that would be the next step. truncating LLM thinking like this does not work well, and its thinking isn't efficient anyway.

teravor · Sep 29, 2026 · comment in “Jeeves. Reasoning improves Jev-like decision models”

Show HN: Saccade – Live semantic browser truth for AI agents I build one this application is to resolve the most of the agent cannot handle the brother they extremely slow.

sacravenger · Aug 31, 2026 · story in “Show HN: Saccade – Live semantic browser truth for AI agents”

I'd like to see real numbers that compare Docker Desktop for macOS before microVMs to post-microVMs. I stopped using Docker on macOS because host file system performance was so slow, even with all of the caching hacks piled on top of it, that it made the whole thing effectively unusable for development.

bradgessler · Aug 10, 2026 · comment in “Docker Sandboxes – Disposable, isolated sandboxes for AI agents”

You have some stuff in WSL that completely are broken once you call a windows native app. Or you have app containers which are too restrictive and result in applying expensive file system ACL to all files the app containers would access, which can take hours when dealing with million of files, and would be required to be applied every time you chains your app container (there are some workarounds, for them but they …

mbrumlow · Aug 3, 2026 · comment in “Qwen3.8-Max: A New Bar for Coding and Cowork”

Coursera stinks now. Extremely slow load times, annoying AI chatbot that takes up the right half of the screen on the landing page, and a lot of courses that were free are now behind a paywall.

mna_ · Aug 3, 2026 · comment in “Coursera Bets $100M That Andrew Ng Can Do What Coursera Can't”

For finding new clusters, sending everything through an LLM was too slow and expensive. HDBSCAN-style embedding clustering also gets painful at scale because of pairwise comparisons.

laalshaitaan · Jul 14, 2026 · story in “Launch HN: Agnost AI (YC S26) – Extract user feedback from agent conversations”

But as long as I still need to review and follow up with changes, Opus is just too slow. It’s really frustrating because it’s a lot slower than it was 6mo ago, and not noticeably better.

anon7000 · Jul 2, 2026 · comment in “CursorBench 3.1”

As there are more and more agents in the internet; Security is going to be a big problem. Currently, the problem is solved using a LLM to guard Agent but this creates the problem of hallucination and latency, so I coded a firewall in rust that runs under five milliseconds.

beebeeVB · Jun 30, 2026 · story in “Show HN: Trajeckt, a Firewall for AI Agents”

As there are more and more agents in the internet; Security is going to be a big problem. Currently, the problem is solved using a LLM to guard Agent but this creates the problem of hallucination and latency, so I coded a firewall in rust that runs under five miliseconds.

beebeeVB · Jun 29, 2026 · story in “Show HN: A Firewall for AI agents with auditing”

Or it takes forever giving my information to a seemingly real person only to be told to wait, just let me fill out a form and give me a queue ETA. And half the time the real agent can't see the AI messages so I'm stuck repeating myself.

wildzzz · Jun 17, 2026 · comment in “Sixty percent of US consumers say 'AI' in brand messaging is a turnoff”

At least codex listens to me telling it to use rg instead of grep, cause grep is often so slow. But when adding rtk it uses grep through rtk which is kind of annoying.

carlmr · May 18, 2026 · comment in “Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep”

Antigravity's workflow is so slow and buggy with permissions, it is unuseable compared to cc/codex. the only part that is nice is that it allows usage of Opus.

jatora · May 8, 2026 · comment in “AlphaEvolve: Gemini-powered coding agent scaling impact across fields”

So I'm feeling pretty validated for that decision these days. The speed improvement was massive (super low latency), and was worth the switch on it's own, but we also saved 90% in immediate cost... probably more in secondary effects from the git host just not being a pain point.

tomxor · Apr 27, 2026 · comment in “GitHub is having issues now”

Git still doesn't work well with non-text data, including being incredibly slow.

bmitc · Apr 13, 2026 · comment in “We've raised $17M to build what comes after Git”

- building and scaling LLM workflows (agents, pipelines, RAG) - handling real-world issues like latency, reliability, and hallucination control

mernaemara · Apr 7, 2026 · comment in “Ask HN: Who is hiring? (April 2026)”

I didn't need anything fancy, just 'did the agent answer the question', 'did the agent cover the 5 things it needed to' - that type of thing. I found myself blocked by 'Gemini in Google Sheets', it was too slow and cumbersome, and it didn't handle eval changes well - particularly when trying to associate evals with ground truth.

raviisoccupied · Mar 29, 2026 · story in “Show HN: Beval – Simple evaluations for your AI product”

Most corporate machines are Windows boxes with ps and cmd.exe heavily restricted, no admin, and anti malware software surveilling I/O like a hawk. You might get a git bash if you are lucky, but it's usually so slow it's completely unusable.

BiteCode_dev · Mar 24, 2026 · comment in “Ripgrep is faster than grep, ag, git grep, ucg, pt, sift (2016)”

If someone knows what I'm doing wrong I'd love to hear it since this would be a great feature. To be honest, I'm honestly sick of Google Home's approach to this since the Gemini update has turned everything really slow and I'm getting close to the point where I'd rather home-roll a full system myself that works reliably instead of the crapshoot that this is.

arjie · Mar 24, 2026 · comment in “Box of Secrets: Discreetly modding an apartment intercom to work with Apple Home”

Ome advantage of python is that it is so slow that if you choose the wrong algorithm or data structure that soon gets obvious. And for complicated stuff this is exactly where I find the LLMs struggle.

tda · Mar 21, 2026 · comment in “We rewrote our Rust WASM parser in TypeScript and it got faster”

The source of latency is calling the Google Gemini API to compute the embedding of the query text. I was hoping to make a demo that searches as you type, but the two second delay makes it more annoying than useful.

jiggawatts · Mar 19, 2026 · comment in “Show HN: Hacker News archive (47M+ items, 11.6GB) as Parquet, updated every 5m”

I know this is just one data point, but I don't notice any latency when typing code in VS Code. It takes a while to start up, and that is annoying especially for quick short editing jobs, but other than that I never notice any sluggishness.

roelschroeven · Mar 11, 2026 · comment in “Writing my own text editor, and daily-driving it”

A Python backend acting as a conductor for several models (integrating Veo, Wan, and custom pipelines). The main challenge was getting the inference latency down so the "edit-preview" loop isn't painful.

Nancylily · Mar 5, 2026 · comment in “Show HN: Textideo – Generate video, audio, and 3D assets in one timeline”

AI has made writing code 10x faster, but shipping is still painfully slow. The bottleneck has shifted downstream - to merge conflicts, flaky tests, inconsistent code quality, and dozens of other frictions that drain productivity and morale.

trunksuzieross · Mar 4, 2026 · comment in “Ask HN: Who is hiring? (March 2026)”

It works completely offline and handles Turkish text surprisingly well for a pure retrieval system. It does exactly what I need without the latency, hardware requirements, or hallucinations of running local AI models.

Yigtwx · Feb 27, 2026 · comment in “Show HN: An offline document search engine for my university's messy PDFs”

I pay for the Lumo+ private chat service and it has excellent integrated search, like Gemini or ChatGPT. EDIT: I have also experimented with creating a local search index for the common tech web sites I get information from - this is a pain in the ass to maintain, but offers very low latency to add search context for local model use.

mark_l_watson · Feb 26, 2026 · comment in “How will OpenAI compete?”

tangential, but deploys/builds that involve worktrees happen to neatly sidestep this since then .git is just a pointer to the real one. i use this to avoid having to otherwise prevent docker from wasting time reading the git info into the build context (especially important for latency if feeding local files into a remote image build)

nagaiaida · Feb 23, 2026 · comment in “Git's Magic Files”

Perceptible latency is somewhere between 10 and 100ms. Even if an LLM was hosted in every aws region in the world, latency would likely be annoying if you were expecting near-realtime responses (for example, if you were using an llm as autocomplete while typing).

yunwal · Feb 22, 2026 · comment in “How Taalas “prints” LLM onto a chip?”

Couldn't find anything that gave me low-latency structured JSON state (UI elements, network, errors, logs) across platforms, local-first, without the usual trade-offs. So yeah, I got fed up and built a small local MCP server to solve it for myself.

sebringj · Feb 10, 2026 · story in “Runtime validation is still fucked in AI coding agents”

The biggest risk isn’t “doing AI wrong.” It’s not building AI fluency at all while the rest of the market moves forward. Once AI systems are in production, new problems appear: cost control, reliability, hallucinations, latency, silent regressions.

danelrfoster · Feb 10, 2026 · story in “Why Every Business Must Engage with AI – and How to Do It Right”

In my day to day as a Product Manager working in a team that ships AI products, I often found myself wanting to do 'quick and dirty' LLM based evaluation on conversation transcripts and traces. I found myself blocked by 'Gemini in Google Sheets', it was too slow and cumbersome, and it didn't handle eval changes well.

raviisoccupied · Feb 9, 2026 · comment in “Ask HN: What are you working on? (February 2026)”

AI has made writing code 10x faster, but shipping is still painfully slow. The bottleneck has shifted downstream - to merge conflicts, flaky tests, inconsistent code quality, and dozens of other frictions that drain productivity and morale.

trunksuzieross · Feb 2, 2026 · comment in “Ask HN: Who is hiring? (February 2026)”

> I simply booted up a VM with an H100, ssh’d into it with Cursor, and prompted the agent to set up an inference server that I could ping from my web generation app. What used to take hours or days of painful, slow debugging now takes literally minutes.

westoque · Jan 23, 2026 · comment in “Show HN: isometric.nyc – giant isometric pixel art map of NYC”

Gemini CLI is too slow to be useful, kind of surprised it was even offered and marketed given how painful it is to use.

mythz · Jan 8, 2026 · comment in “How Google got its groove back and edged ahead of OpenAI”

Most of my direct code editing has been in the form of defining consts and const data structures, then telling Gemini how I want those used (rules, game logic). I have done virtually all of the (play)testing, because at this point I find the Antigravity web browser driver experience way too slow and unreliable.

michaelteter · Jan 5, 2026 · story in “Show HN: Vector Defense – retro tower defense game”

AI has made writing code 10x faster, but shipping is still painfully slow. The bottleneck has shifted downstream - to merge conflicts, flaky tests, inconsistent code quality, and dozens of other frictions that drain productivity and morale.

trunksuzieross · Jan 2, 2026 · comment in “Ask HN: Who is hiring? (January 2026)”

You talk to an AI that goes incredibly slow and tries to get you to add extras to your order. I would say it has made the experience more annoying for me personally.

lutharvaughn · Dec 26, 2025 · comment in “Rob Pike goes nuclear over GenAI”

Both have near-conversational latency. In my own agent I struggle to get sub-second, even with OpenAI streaming for TTS.

Norcim133 · Dec 26, 2025 · story in “Ask HN: How does Boardy achieve such low latency?”

I dislike the official git completion script because it’s so slow when working with large repos. On macOS - because of vagaries of the way its file system works - with large repos it’s effectively unusable.

derriz · Dec 25, 2025 · comment in “Archiving Git branches as tags”

It doesn't have to be very intelligent. When I tried gpt-oss and qwen using ollama on an M2 Mac the main problem was that they were extremely slow.

Myrmornis · Dec 22, 2025 · comment in “A guide to local coding models”

- Voice Agents: Shaving 200ms off RAG latency to make conversation feel natural. - Multi-Agent Hand-offs: Agent A (Sales) updates a "Customer Mood" state, and Agent B (Support) sees it instantly without hallucinating.

aperi · Dec 9, 2025 · comment in “O(1) Context Retrieval for Agents Using Weightless Neural Networks”

Show HN: Bifrost – open-source LLM Gateway (50x lower latency than LiteLLM) We built Bifrost because we found existing Python-based gateways struggled with high concurrency in production.

dskuldeep · Dec 9, 2025 · story in “Show HN: Bifrost – open-source LLM Gateway (50x lower latency than LiteLLM)”

AI has made writing code 10x faster, but shipping is still painfully slow. The bottleneck has shifted downstream - to merge conflicts, flaky tests, inconsistent code quality, and dozens of other frictions that drain productivity and morale.

trunksuzieross · Dec 1, 2025 · comment in “Ask HN: Who is hiring? (December 2025)”

You're burying the lede: SOTA 'Reasoning Models' (o1/GPT-4) are actually unusable for agent swarms because inference latency kills the recursion loop.

Shahaf_Wieder · Nov 24, 2025 · comment in “BYO – A multi-agent runtime optimized for parallel inference”

Careforce AI | Senior Fullstack Engineer | REMOTE (US) | Full-time Healthcare administration is stuck in the era of the fax machine—manual, frustrating, and incredibly slow.

jackhelpcareai · Nov 4, 2025 · comment in “Ask HN: Who is hiring? (November 2025)”