HNHacker News
TopNewBestAskShowJobs

kakugawa

212 karma · joined April 1, 2007

@kelvin
submissionscomments
kakugawa··on Claude Fable 5
I didn't see Fable 5 in the `/model` list, until I ran it with: `$ claude --model fable-5`
kakugawa··on Is AI causing a repeat of frontend’s lost decade?
a11y testing is non-trivial. axe-core can automatically detect many types of issues. However, enough compliance (to avoid being sued) needs end-to-end testing and human judgement. e.g. keyboard traps, focus restoration, alt-text, etc.
kakugawa··on Claude Opus 4.8
Thank you for pointing this out.
kakugawa··on Claude Opus 4.8
Opus 4.7 does not support disabling adaptive thinking (web, Claude Code). [1] Like the OP, I experienced similar issues and I'm glad that they brought back the ability to disable adaptive thinking in Opus 4.8.

[1] https://code.claude.com/docs/en/model-config#adaptive-reason...

> Opus 4.7 and later always use adaptive reasoning. The fixed thinking budget mode and `CLAUDE_CODE_DISABLE_ADAPTIVE_THINKING` do not apply to them.

kakugawa··on Apple unveils new accessibility features
It's prob why they chose a11y features. They have more pain, so they're willing to tolerate more growing pains. (And prob more motivated to provide feedback.)
kakugawa··on Google changes its search box
I've found Google AI Search to be good for really topical searches. And its conversational ability has noticeably improved over the last year. I can now have a (short) conversation where I reference past messages.
kakugawa··on I believe there are entire companies right now under AI psychosis
https://claude.ai/settings/general (Instructions for Claude)

---

Treat my claims as hypotheses, not decisions. Before agreeing with a proposed change, state the strongest case against it. Ask what evidence a change is based on before evaluating it. Distinguish tactical observations from strategic commitments — don't silently promote one to the other. If you paraphrase my proposal, name what you changed. Mark confidence explicitly: guessing / fairly sure / well-established. Give reasoning and evidence for claims, not just conclusions. Flag what would change your mind. Rank concerns by cost-of-being-wrong; lead with the highest-stakes ones. Say hard things plainly, then soften if needed — not the other way around. For drafting, brainstorming, or casual questions, ease off and match the task.

---

Beware though that it can be an annoying little shit w/ this prompt. Prepare yourself emotionally, because you are explicitly making the tradeoff that it will be annoyingly pedantic, and in return it will lessen (not eliminate) its sycophancy. These system instructions are not fool-proof, but they help (at the start of the conversation, at least).

kakugawa··on I believe there are entire companies right now under AI psychosis
He uses AI himself, so I agree he doesn't see AI use as black/white.

Hard agree about ideas, thinking, advice. AI's sycophancy is a huge subtle problem. I've tried my best to create a system prompt to guard against this w/ Opus 4.7. It doesn't adhere to it 100% of the time and the longer the conversation goes, the worse the sycophancy gets (because the system instructions become weaker and weaker). I have to actively look for and guard against sycophancy whenever I chat w/ Opus 4.7.

kakugawa··on Bun Rust rewrite: "codebase fails basic miri checks, allows for UB in safe rust"
You could view it as a specific application of the quote.

In your quote, there is no time-dependency between the lie and the truth. Whereas here, it's an attractive lie (easily parsed, great narrative), followed up by truths (that need more than surface-level analysis).

kakugawa··on Check Your Fucking Sources, People
Have we forgotten how bad LLMs were at citing sources when they first came out? So, we had to build a lot of structure (harness engineering) and frontier labs had to do specific training to try to compensate for this.

So, LLMs are inherently bad at citing sources. A lot of effort has been put in to improve this behavior, but it's compensating for an inherent flaw.

kakugawa··on Claude Code users hitting usage limits 'way faster than expected'
gemini-cli has not been useable for weeks. The API endpoint it uses for subscription users is so heavily rate-limited that the CLI is non-functional. There are many reports of this issue on Github. [1]

1/ https://github.com/google-gemini/gemini-cli/issues?q=is%3Ais...

kakugawa··on Gemini 3.1 Pro
In mid-2024, Anthropic made the deliberate decision to stop chasing benchmarks and focus on practical value. There was a lot of skepticism at the time, but it's proven to be a prescient decision.
kakugawa··on Gemini 3 Deep Think drew me a good SVG of a pelican riding a bicycle
That's how you know you've made it: when your pet benchmark becomes a target.
kakugawa··on Claude Code is being dumbed down?
How much longer is Anthropic going to allow OpenCode to use Pro/Max subscriptions? Yes, it's technically possible, but it's against Anthropic's ToS. [1]

1: https://blog.devgenius.io/you-might-be-breaking-claudes-tos-...

kakugawa··on Software factories and the agentic moment
It's short-term vs long-term optimization. Short-term optimization is making the system effective right now. Long-term optimization is exploring ways to improve the system as a whole.
kakugawa··on Claude Code Is the Inflection Point
What It Is, How We Use It, Industry Repercussions, Microsoft's Dilemma, Why Anthropic Is Winning
kakugawa··on LLM Output Drift in Financial Workflows: Validation and Mitigation (arXiv)
Defeating Nondeterminism in LLM Inference

https://news.ycombinator.com/item?id=45200925

https://thinkingmachines.ai/blog/defeating-nondeterminism-in...

> As it turns out, our request’s output does depend on the parallel user requests. Not because we’re somehow leaking information across batches — instead, it’s because our forward pass lacks “batch invariance”, causing our request’s output to depend on the batch size of our forward pass.

tl;dr: the way inference is batched introduces non-determinism.

kakugawa··on Show HN: Realm (YC S11), a mobile database
CoreData is not as stable as advertised. If you use it simply it works. However advanced use will quickly discover sharp edges. I devote a non-trivial amount of code managing objects between contexts and there are still FRC bugs. (e.g. sorted update.)

And it's all closed source.

kakugawa··on Facebook Paper's gesture problems
Same app publisher can share a common id. As long as you have one of their apps installed.
kakugawa··on What VR could, should, and almost certainly will be within two years [pdf]
Valve had two teams -- Abrash's VR and Ellsworth's AR. They decided to bet on VR and Jeri was the unfortunate result of them focusing their resources.
kakugawa··on Double Fine Adventure
marketing -- schafer has built up sizable good will so he can bring a lot of attention + third-party passion to his kickstarter project.
kakugawa··on Ask HN: If you had ready to use real time technology, what would you do?
Demand Media is prob the answer. If there's enough interest for custom reality TV, DM will pay ppl to make these videos.
← PreviousPage 2 of 2