waldrews

10 ms·

1.
by waldrews 4d ago
Yup. The problem is that it's bizarrely still not in General Availability status.
2.
by waldrews 4d ago
Completely free flow typing is risky in terms of interpretability, but type narrowing - var a : supertype; if (a is subtype) { // a is known to be subtype }, or type case, saves boilerplate in any OOP language.
3.
by waldrews 4d ago
Yup, tried all variations of that. There's the advantage that you can use a lower hallucination OCR specific model for the pre-processing, at least for clean text. But for something hard like handwritten forms, applying VLM with cont
4.
by waldrews 4d ago
We sure did. It's a great writer, better in a harness, will process lots large context, but complex reasoning with convoluted rules and low hallucination tolerance? That's still larger model territory.
5.
by waldrews 5d ago
It's still 'preview' and not generally available, so can't run it for US restricted workloads.
6.
Ask HN: Did Google kill its enterprise workhorse model?
27 points by waldrews 5d ago|18 comments
7.
by waldrews 15d ago
(off topic) why the headache-inducing animated background? An annoyance for all, and an actual accessibility issue for some.
8.
by waldrews 21d ago
Every technical topic involves AI now. Even if it's about why the ancient Greeks didn't turn their steam engines to industrial use - the AI's should be researching the details. We might as well ask which topics involve huma
9.
by waldrews 1mo ago
TreasuryDirect's login and account recovery experience has been notorious for years, both for user experience and for people easily getting locked out for weeks. It's good they're being careful with this rollout, as it serve
10.
by waldrews 1mo ago
The VLM's are so good at complex document understanding now. But you just can't trust them not to invisibly censor sensitive clinical/legal docs, even at the maximally permissive settings. And the deep learning OCR-only mode
11.
by waldrews 1mo ago
The only mainstream-ish language where that still happens? R. Too bad, there's a lot more use cases for this sort of thing now - versioning, anything non-persistent agents touch, collaboration, auditable enterprise LOB. It's 20
12.
by waldrews 2mo ago
3.5 Flash-Lite seems available in US region, as was 3.5 Flash; but 3.6 Flash looks Global only so far when pinging. If Google employees are watching, will this issue go away?
13.
by waldrews 2mo ago
Very farsighted, after working as a patent clerk, to lay claim on such a foundational technology. Back in the day, they must've been like, oh, so Mercury blocks the sun at the wrong time, but where's the commercial value - and no
14.
by waldrews 3mo ago
Back in my day, a software company of national importance would be considered a big success if acquired for 60 million... Of course, at that point it would have to be profitable, and actually own its key IP...
15.
by waldrews 5mo ago
Not quite the same scenario, but it's already plausible to have a situation where every subagent is allowed to spawn multiple subagents, in which case we'd have literally exponential credit consumption growth...
16.
by waldrews 5mo ago
Startup working on AI Healthcare solutions (plus some random experiments) | REMOTE(US, Pacific time zone) | Full time and project based Full time engineering role (Remote, US person required; Pacific time zone) Stack: C#, Python, Gemini AP
17.
by waldrews 5mo ago
File systems are nice if you need to do manual or transparent script-based manipulations. Like 'oh hey, I just want to duplicate this entry and hand-modify it, and put these others in an archive.' Or use your OS's access co
18.
by waldrews 6mo ago
Aargh, aggressively blinking visual horror website.
19.
by waldrews 7mo ago
The new workflow will be "AI, I need to view this text file and add some words to it. Create an app that displays it in a scrollable window, respecting the encoding. Now move the cursor to the line below the three dashes... no, the o
20.
by waldrews 7mo ago
well, sure, that uses a large number of processing cycles for each small operation. But asking a frontier LLM to evaluate a lisp expression is more or less on the same scale (interesting empirical question whether it's more or less).
21.
by waldrews 8mo ago
And they're pretty much the only example of an embedded browser architecture actually performing tolerably and integrating well with the native environment.
22.
by waldrews 8mo ago
That's one maxed out RAM configuration. Back in my day, we had 4k RAM, about 3500 bytes usable from BASIC, and that was enough, unless you were rich enough to have a 3k memory expansion cartridge. But really, if you need that extra 3
23.
by waldrews 8mo ago
We're just not going to see any code written entirely without AI except in specialist niches, just as we don't see handwritten assembly and binaries. So the disclosure part is going to become boilerplate. In the old era, the comb
24.
by waldrews 8mo ago
There's a ridiculous amount of tech in the DNA and cellular machinery of a single bacterium.
25.
by waldrews 8mo ago
My whole cat is, what, a couple gigs of DNA?
26.
by waldrews 8mo ago
Yup! My point is that the 'coin flip baseline' model that's as good as chance isn't actually trivial to create, for an unbalanced and time varying underlying distribution.
27.
by waldrews 8mo ago
The threshold isn't 50% because the distribution of human and AI written cases isn't naturally 50-50. So a coin flip will underperform always guessing the more frequent class. Where it gets interesting is if the base is unknown
28.
Ask HN: Who's given up on getting hired? (January 2026)
8 points by waldrews 9mo ago|1 comments
29.
by waldrews 9mo ago
Why didn't they just ask ChatGPT? Oh wait.
30.
by waldrews 9mo ago
I'm being hyperbolic of course, but I'm a little dismissive of the progress that happened since the days of BBS's and car based cell phones - we just got more connectivity, more capacity, more content, bigger/faster. Li

More ›