a_wild_dandan

6 ms·

31.
by a_wild_dandan 1y ago
Our modern world was built for hand-havers, so we build hand-havers for general interaction with the modern world. This vicious cycle will end, but a certain amount of inertial clunkiness is inevitable in nascent technological disruptions.
32.
by a_wild_dandan 1y ago
I blame my tool, Cursor. They blame their tool, open-vsx. We're either both right about that logic, or both wrong. Either way, I expect consistency in how the product I pay for assigns/accepts blame. Cursor's response will be
33.
by a_wild_dandan 1y ago
That's probably why the old saw isn't just "put your money."
34.
by a_wild_dandan 1y ago
> The other inventions would have quite the adoption rate if they were similarly subsidized as current AI offerings. No, they wouldn't. The '80s saw obscene investment in AI (then "expert systems") and yet nobody&#x27
35.
by a_wild_dandan 1y ago
This is a fantastic framing method. Anyone who sees the future differently to you can be brushed aside as “an inevitablist,” and the only conversations worth engaging are those that already accept your premise. --- This argument so easily c
36.
by a_wild_dandan 1y ago
I'm uncertain if MY code works correctly lol. I know many code-illiterate folk; some of them I call "boss" or "client." They get along fine dining on my spaghetti. I do likewise never touching the wheel/pedal
37.
by a_wild_dandan 1y ago
Altman released GPT-2 despite expressing that doing so was a bad idea? That's wild.
38.
by a_wild_dandan 1y ago
Local-first doesn't mean local-only though, yeah? Isolate cloud usage to those collaborative features. If that's a huge ask, then your thingy probably isn't the kind of tool we're talking about localizing here!
39.
by a_wild_dandan 1y ago
What does that matter? We're talking trifectas here, not supermajorities. The filibuster is a cute remnant of "decorum." It's a vestigial rule which will disappear when too inconvenient. (Fun question with not-so-fun ans
40.
by a_wild_dandan 1y ago
I purchased "approximately $0.00" in TF2 loot boxes. How much exactly? Left as an exercise to the reader.
41.
by a_wild_dandan 1y ago
Consumer printers work well (or even better) without any dampers, and they come in all (e.g. fishbowl) sizes.
42.
by a_wild_dandan 1y ago
I love this retort and I'm stealing it.
43.
by a_wild_dandan 1y ago
Content authenticity follows from the source's credibility. This is why chains of evidence are crucial. If squinting at pixels were the harbinger of a post-truth reality, then it's already many decades late.
44.
by a_wild_dandan 1y ago
I think this says more about Scottish than the model.
45.
by a_wild_dandan 1y ago
Perhaps Starfield was the most important Bethesda release. The animus toward Starfield will serve as an enormous signal/reminder to course correct away from this "unique feature." One can hope.
46.
by a_wild_dandan 1y ago
Which developer jobs aren't dependent on many third parties?
47.
by a_wild_dandan 1y ago
That's using a mighty broad brush to paint folks' circumstantial situations.
48.
by a_wild_dandan 1y ago
Remembers everything that you say, isn't limited to an hour session, won't ruin your life if you accidentally admit something vulnerable regarding self-harm, doesn't cost hundreds of dollars per month, etc. Healthcare is abou
49.
by a_wild_dandan 2y ago
I suppose Google wants us to pretend that "agents" can't be "resources." MCP is already well established (Anthropic, OpenAI, Cursor, etc), so Google plastering their announcement with A2A endorsements just reeks of
50.
by a_wild_dandan 2y ago
Exposing invalid transitions to a user is a bug. Idempotency here doesn't solve anything, just hides said bug, which is arguably worse .
51.
by a_wild_dandan 2y ago
When you identify where the infringing party has stored the source material in their artifact.{zip,pdf,safetensor,connectome,etc}. In ML, this discovery stage is called "mechanistic interpretability", and in humans it's calle
52.
by a_wild_dandan 2y ago
GLP-1 drugs offer palliative mercy for terminal societal malaise. You're in the bargaining stage.
53.
by a_wild_dandan 2y ago
And for anyone looking to dig deeper, check out "grammar-based sampling."
54.
by a_wild_dandan 2y ago
> We evaluate Qwen2.5-Max alongside leading models > [...] we are unable to access the proprietary models such as GPT-4o and Claude-3.5-Sonnet. Therefore, we evaluate Qwen2.5-Max against DeepSeek V3 "We'll compare our propri
55.
by a_wild_dandan 2y ago
If I replaced "transformer" in your comment with "human", what changes? That's my point. Humans are a "function of historical data" (nurture). Meatbag I/O doesn't span all sequences. A person&#
56.
by a_wild_dandan 2y ago
> [...] this technology which has progressed only by increasing data volume and variety Sure, if you ignore major shifts after 2022, I guess? Test-time-compute, quantization, multimodality, RAG, distillation, unsupervised RL, state-space
57.
by a_wild_dandan 2y ago
Running a 680-billion parameter frontier model on a few Macs (at 13 tok/s!) is nuts. That'a two years after ChatGPT was released. That rate of progress just blows my mind.
58.
by a_wild_dandan 2y ago
People will downvote you for revealing this, but it's the truth. I saw it on TikTok, after all.
59.
by a_wild_dandan 2y ago
From TFA/paper: "We conduct a variety of experiments to demonstrate that our results cannot be exclusively due to implicitly priming the model to fake alignment, as we overview in Section 8.1. For example, we find that alignment f
60.
by a_wild_dandan 2y ago
friction

More ›