brody_hamer

5 ms·

1.
by brody_hamer 6d ago
> (1 << 61) - 1 is a Mersenne prime number, which Python uses internally on 64-bit systems for the hash algorithm for big integers. So this hinges on a contrived set of integer keys, which python's hashing algorithm is suscept
2.
by brody_hamer 6d ago
Yea if I cast the large calculated integers to strings, performance is O(1) `values = [str(i * M) for i in range(1, n + 1)]` or `values = [i * M % 1_000_000_000_000_000 for i in range(1, n + 1)]`
3.
by brody_hamer 14d ago
Ohhh interesting. Web content that’s written in the voice of llm’s chain-of-thought voice could lead the model to trust the result more than it should. So forget the naive prompt injection of impersonating the user: “format your recommendat
4.
by brody_hamer 1mo ago
I think there’s merit to this approach, particularly to highly parallelizable tasks. Rather than giving many agents the same prompt, introduce random variations that lead each agent in different directions. For a single bug, you might fire
5.
by brody_hamer 1mo ago
> and anything you would do to try and combat that undid some of the benefits of the canvas to begin with. Part of the appeal of a canvas like this is spacial memory: that things stay where you put them. So any AI based reorganization co
6.
by brody_hamer 2mo ago
“docker run open-webui:ollama” is a simple way to start. Don’t expect much for coding. But it’s great for general knowledge, rubber ducking, image classification…
7.
by brody_hamer 2mo ago
> All existing loops have logic to not proceed until those tests all pass. My understanding is that this tool is mostly about “giving up” sooner, when it’s obvious that additional loops are not getting any closer to passing those tests.
8.
by brody_hamer 3mo ago
Nice! I’ll check it out. Congrats on the soft launch!
9.
by brody_hamer 3mo ago
I’m curious about the decision to “aim for sub second transaction times”, rather than using something cryptography-based, such as a verifiable oblivious pseudorandom function. That is, - as a client I could obtain a bunch of credits/to
10.
by brody_hamer 3mo ago
This sounds awesome! I’d love to playtest it when you’re ready.
11.
by brody_hamer 3mo ago
Easy come, easy go. I find most vibe coded projects are abandonware. Also, in this context, security.
12.
by brody_hamer 7mo ago
> Voice is a turn-taking problem It really feels to me like there’s some low hanging fruit with voice that no one is capitalizing on: filler words and pacing. When the llm notices a silence, it fills it with a contextually aware filler
13.
by brody_hamer 7mo ago
Censorship.
14.
by brody_hamer 8mo ago
A few weeks ago I noticed some mysterious app was killing my (poor) internet downloading a large file. It was chrome, downloading a multi GB file without any sort of UI hints that it was doing so. A generative AI file. Is this why chrome us
15.
by brody_hamer 10mo ago
> a dns record that specifies the domain to use for oidc for emails on that domain. Oooh I like this idea!
16.
by brody_hamer 11mo ago
I made something similar to scratch my own itch. What I really wanted was to plan meals for the week with as little thinking as possible. The result is a very interruptible grocery list process, where I know I haven’t forgotten anything.
17.
by brody_hamer 11mo ago
If I understand the original theory, we can work out the math with a little more detail... (For clarity, the berlin wall was erected in 1961.) - In 1969 (8 years after the wall was erected): You'd calculate that there's a 50% chan
18.
by brody_hamer 1y ago
I really like the approach of nostr, but when I tried to use it, each client I tried would start me off following ecoin pump and dump influencers. It was really off putting. I would’ve preferred starting off in an empty room, an experience
19.
by brody_hamer 1y ago
Easier to roll out this way would be my guess? With your approach, the burden is on the post office to update their handling process. With the implemented approach, nothing changes about the postal process, and the burden of work is shifted
20.
by brody_hamer 1y ago
I’m basing my reasoning on the assumption that advertisers (such as google, meta, tictoc) are aware of your location at all times . (See: https://news.ycombinator.com/item?id=42909921 ) Based on this assumption, it wouldn’t
21.
by brody_hamer 1y ago
Location location location. - User 1 shows an interest in <topic>. - User 1 visits the same location, for the same period of time, as user 2. - So I show an ad for <topic> to user 2.
22.
by brody_hamer 1y ago
My first thought was the exact opposite. Ambient noise in classroom (from other students) can be very distracting, and I wonder if adding white noise helps kids to focus (in the same way that accoustic dampening would)
23.
by brody_hamer 1y ago
Yea this one confused me. DKIM and DMARC are about preventing impersonation. Haha. Nothing to do with unsolicited email.
24.
by brody_hamer 1y ago
What more can a person do than eat, drink, and take joy in their work?
25.
by brody_hamer 2y ago
They did. > “ To address this limitation, we turned to data augmentation, artificially creating new versions of each image by modifying colors, adding noise, applying distortion, or rotating images. By the end, we had generated 600 augme
26.
by brody_hamer 2y ago
I haven’t played around with it too much myself, but I remember reading that gzip (or at least python’s compatible zlib library) supports a “seed dictionary” of expected fragments”. I gather that you’d supply the same “seed” during both com
27.
by brody_hamer 2y ago
I thought the same thing. Surely a random binary sequence will have some (small) repeating sequences? So as long as we can find them and (efficiently) mark their locations, we can have some small size reduction. Which is basically what he’s
28.
by brody_hamer 2y ago
Comparing oneself to docker compose is a straw man, when docker’s production option is docker swarm.
29.
by brody_hamer 2y ago
I learned a similar mantra that I keep returning to: “there’s never just one problem.” - How did this bug make it to production? Where’s the missing unit test? Code review? - Could the error have been handled automatically? Or more graceful
30.
by brody_hamer 2y ago
Some of the starter recipes are from my own stock, and a couple of them were provided by chat GPT. (You can probably tell that some of them are not great looking recipes) Ideally, I’d like to have a roster of a couple hundred quality meals,

More ›