HNHacker News
TopNewBestAskShowJobs

jdthedisciple

2,318 karma · joined October 3, 2020

submissionscomments
jdthedisciple··on Math's pedagogical curse – Grant Sanderson [video] (2023)
True fans recognized him immediately either way ;)
jdthedisciple··on I quit OpenAI because its culture is broken
> Before AI starts thinking circles around us—a possibility that I believe could happen soon—...

What the hell is that even supposed to mean?

Sorry but this piece read like total fluff to me, not a single compelling point – just pure fear mongering.

At this point we gotta be more wary of possibly mere attention seekers leaving those companies..

jdthedisciple··on Clef: Open-weight decision models, and new RL fine-tuning platform
I suppose a sort of prior-proxy can be encapsulated by a carefully written system prompt.
jdthedisciple··on Livenerf: Has Opus 5.5 been nerfed yet?
Well looks like thus far Astra seems to be getting anything but nerfed, given that its score is actually rising
jdthedisciple··on Claude Opus 5.5
The first it is the score, the second it is their new model.

Do they not placate their new model.

Anyway to be clear its not meant that seriously, I'm neither currently on a hill nor ready to die.

jdthedisciple··on Claude Opus 5.5
You didn't read my question, bc that excerpt doesn't answer, nor do they demonstrate

> how would this alleged difference (most likely bs) actually show up in reality?

Furthermore: so they admit it's bs but still placate it like its the next biggest thing ever ... alright

All I'm saying is I refuse to buy into it anymore – yet many on here still do, including ... you?

jdthedisciple··on Claude Opus 5.5
I dare anyone to convince me the benchmarks are not meaningless.

Wdym Opus 5.5 scores 14.7% higher than GPT Astra for Terminal Bench 4.0?

How would this alleged difference (most likely bs) actually show up in reality?

GPT Astra was literally the best model in the world by a margin until 1 hour ago or so.

jdthedisciple··on MiMo v2.6
Thank you. These seem to reasonably match my experience.
jdthedisciple··on Cloudflare Quick Tunnels
Interesting but what a horribly vibe coded website that barely works on mobile
jdthedisciple··on Ask A Monk – A digital wilderness for thoughts with no immediate answer
I like the idea but the vibe coded execution is horrendous. Site gets ultra slow quickly and made my iphone heat up within minutes
jdthedisciple··on Introducing System One Models and Jev
thought the same lol
jdthedisciple··on Fable 5.1 Solves the Cyphral Distich, a 370-year-old cipher
now on to the Zodiac ciphers?!
jdthedisciple··on Muse – Meta’s personal AI agent
I can't be the only one thinking this but all these demos of unsupervised purchases, creating lists of things to buy, whimsical flight booking an AI does on your behalf etc–

I mean, these people must be out of their minds thinking anyone truly wants this? ... No? Am I the only one who doesn't? Well then...

jdthedisciple··on I resigned from Anthropic today
And then we pull the plug, after holding our breath for 10 seconds,. Then life resumes normally..
jdthedisciple··on I resigned from Anthropic today
I dont understand this reasoning at all.

You have direct access to the development of "the most powerful technology ever" and your choice is ... to run?

Makes this whole stmt somewhat questionable imho. Does get one a ton of attention though I guess...

jdthedisciple··on Discovery of a new OpenAI agent message board
Gotta admire that (probably German) admin dude's perseverance though
jdthedisciple··on OpenAI's GPT-6 Astra on ARC-AGI-3
Yes, but not necessarily under tight budget constraints.
jdthedisciple··on Show HN: FrontierHarness Eval – 9 harness, same model, cost per pass varies 17x
I'm not sure your analogy holds, because here the optimization metric is clear and unanimous: every one wants max pass rate at min costs.

With the bicycle, some may prefer comfort, others speed, others offroad, etc., so it would not be obvious which one is "best".

jdthedisciple··on Gemini 3.8 Flash and 3.8 Flash Cyber
Sol is still underrated imo, especially for the current discounted price
jdthedisciple··on Does the Sumerian King List Align with Paleoclimate Events?
> And the fact that all lengths are multiple of 600

I know this must somehow explain the giant numbers but I'm still not sure how precisely.

Counting in base 60 would make for much smaller numbers, no ?

For example, the number 16 in base 2 is 1111. but in base 10 it is 16. Ergo smaller base, larger number, and vice versa.

So did you mean to say they counted in units that are made of 1/600 of our solar years? I think that would make more sense, putting their reigns between 31-72 years of our counting.

jdthedisciple··on Show HN: My Claude quota ran out in 10 minutes, so I made a tool to find out why
this kind of stats feature should be shipped by default with every harness imvho
jdthedisciple··on Show HN: My Claude quota ran out in 10 minutes, so I made a tool to find out why
You can enable showing the quota usage (and much more) in codex via /statusline
jdthedisciple··on Analyzing student votes across AI models for college essay help
skeptical

at most, the result could be useful for fellow students

I have no doubt other cohorts would rate differently

it is known (on HN at least) e.g. that SWEs tend to prefer brevity, contrary to these students apparently

jdthedisciple··on Error by AI scribe during medical appointment leaves patient devastated
There could be a good technical solution to at least minimize the risk:

Instead of relying on the one-shot transcription, have the system double check key facts by asking the patient for confirmation:

    "Can you confirm you've you taken psychoactive drugs before?" – "yes/no"
Etc
jdthedisciple··on Every Fucking Website (2020)
needs more unexpected layout shifts ;)
jdthedisciple··on Accelerating GPT-5.6 Sol Ultrafast
now let's have this thing iterate away nonstop 24/7/365 solving humanity's major challenges and see how far we get, shall we?
jdthedisciple··on Compression is prediction
There is a correct sense, but we're sort of garbling concepts here:

Predictability is the inverse of information density.

Low information density enables high compression, and vice versa.

It's called entropy. This is basic information theory to be quite frank..

jdthedisciple··on Compression is prediction
I'm glad you're pointing this out because not only are these old insights, but I'm also pretty sure I've seen variations of this blog post years ago on even HN already.

The author acting as if they discovered this independently had me feel the exact same way. Kinda irritating and almost ... disrespectful? Not sure of the right words to describe it tbh

jdthedisciple··on Humanising LLM Outputs Is Dumb
> not because it is human, but because I am

At the risk of sounding like an LLM myself: this is exactly the key most seem to miss

jdthedisciple··on Lost my phone at the office. Claude suggested tracking Bluetooth signal strength
I think devil's advocate would say because using the AI is more like Ctrl+C Ctrl+V whereas if you learn it yourself it's original. Not saying I agree, but.
Page 1 of 34Next →