kakugawa

8 ms·

1.
by kakugawa 6d ago
I assume you'd develop via the SDK, then deploy it via the API.
2.
by kakugawa 11d ago
It's not a coincidence that Waymo started becoming viable after GPT-3.
3.
by kakugawa 14d ago
With all the hype around the latest Gemini 3.8 Flash/Cyber release, will Antigravity CLI [1] be supported? 1/ https://antigravity.google/product/antigravity-cli
4.
by kakugawa 16d ago
https://www.youtube.com/watch?v=c2yCePPnrSA Their analogy is that a VLM responds to text, like their interface model responds to clicks. There is no UI (just images), and based on your clicks the model infers your intent an
5.
by kakugawa 19d ago
https://autosaddler-projectpage.github.io/ https://github.com/microsoft/AutoSaddler
6.
by kakugawa 20d ago
They make money on enterprise plans: https://www.experientiallabs.ai/pricing#enterprise Look at the Intelligence features in the Enterprise plan: * Per-prompt model optimization * Caching * A model you own, trained on your
7.
by kakugawa 20d ago
FrontierCode is prob the closest. [1] It's closed source (so no direct benchmaxxing), and it was calibrated by 20+ open source maintainers. It shows Opus 5 (medium), beating out the other reasoning levels by a large margin. i.e. Opus 5
8.
by kakugawa 27d ago
https://archive.is/iV3sk
9.
by kakugawa 1mo ago
I believe Entire.io is also trying to build a Github replacement. (Not affiliated, but I use the Entire.io CLI and I find it useful.)
10.
by kakugawa 1mo ago
"Advanced" can just mean that agents perform actions at a high enough velocity that a human operator can't reasonably review it. i.e. what is already possible today.
11.
by kakugawa 1mo ago
I feel like there's going to be a migration to a new interface paradigm that's balanced between agentic use and human review (with light editing). The current low/no code UIs are human-centric (visual, point-and-click). Going
12.
by kakugawa 2mo ago
They can at least deliver baseline value, because it's all text-based. Do agree that it'll become more complex, when/if they decide to create custom protocols.
13.
by kakugawa 2mo ago
The prefix cache is a resource shared by all users. This is basically a tragedy of the commons.
14.
by kakugawa 2mo ago
The more subtle point is that there's a gap between the task and its verification. e.g. if you have an open-ended / under-specified prompt, the verification needs to be able to handle all potential solutions. So you can have a ver
15.
by kakugawa 3mo ago
Antigravity CLI (which replaced Gemini CLI): https://antigravity.google/product/antigravity-cli
16.
by kakugawa 3mo ago
Here's the demo: https://docs.qwenlm.ai/resources/mlu56_demo.html Here's the description of the world model prompt for the web domain: "A precise GUI state simulator — given the current screen (as HTML)
17.
by kakugawa 3mo ago
How much does /goal actually help? In auto mode, I've tried using and not using /goal and I haven't felt a difference. https://code.claude.com/docs/en/goal#how-evaluation-works > /goal
18.
by kakugawa 3mo ago
Are the "clicks around the UI" converted into end-to-end tests eventually? e.g. via playwright.
19.
by kakugawa 3mo ago
So, how is it being disabled? It still shows "Fable 5" on all surfaces (to me). Is it being silently degraded to Opus under-the-hood? Edit: Fable 5 was just disabled.
20.
by kakugawa 3mo ago
I didn't see Fable 5 in the `/model` list, until I ran it with: `$ claude --model fable-5`
21.
by kakugawa 4mo ago
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.
22.
by kakugawa 4mo ago
Thank you for pointing this out.
23.
by kakugawa 4mo ago
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://co
24.
by kakugawa 4mo ago
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.)
25.
by kakugawa 4mo ago
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.
26.
by kakugawa 4mo ago
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
27.
by kakugawa 4mo ago
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
28.
by kakugawa 4mo ago
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
29.
by kakugawa 4mo ago
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 inherentl
30.
by kakugawa 6mo ago
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&

More ›