Hacker Newsnew | past | comments | ask | show | jobs | submit | sireat's commentslogin

As I recall Holmes did in fact do a lot of walking. He vacillated between periods of inactivity(cocaine, violin, shooting V in wall with a revolver) and intense activity (taking up disguises and doing various physical activities including walking all across London and elsewhere.

Just because your logical mind says one thing is good to do and you know you should do it you are not going to always obey your rider, the inertia of the elephant takes over.

So you need a trigger to snap out of it, for Holmes it was a new case.


> and intense activity

AFAIR those had a specific purpose (chasing a perp, tracking down evidence, etc.). Most of his thinking he did sitting in a chair and smoking his pipe for hours on end (sometimes the whole night).


No they are not. He plays violin and shoots a gun inside his house for fun.

Indeed regular Jupyter works so well on VS Code for solo work these days that there is no real need for a new entrant.

So what pain point are these new entrants trying to solve?

Sure there is an issue of .ipynb basically being a gnarly json ill suited for git but it is rare that I need to track down a particular git commit. Even then that json is not that hard to read.

Also I'd like an easier way to copy cells across different Jupyter notebooks, but at the end of day it is just Python and markdown not very hard to grok.


OpenAI has ridiculous guardrails for illustrations covering any public domain subject that has been covered by Disney or any other major public corporation.

So by that benchmark Japanese companies have a case.

Try generating a 19th century style illustration of Snow White. You can't at least not on OpenAI platform.

Try generating a picture "of flying boy fighting a pirate on a ship".


I have a story from the mechanical side.

I spent a month in 2012 roughly 4 hours a day doing various tasks.

It was horrible, even if I followed all the "best practices" of Turkers it was not a way to make a living.

By end of the month, I had become so jaded to all the "priming" experiments by graduate and undergraduate psychology students. Those usually paid at least something 3-4 USD an hour.

Did some porn labeling tasks, those were horrible after the novelty wore off.

Did very few other labeling tasks because they paid next to nothing.

To have someone actually depend on living for these seemed like a torture.


What country do you live in?

There are places where $3-$4 USD per hour is significantly higher than the prevailing wage. This is not a great fact about global wealth disparity, but that money goes towards improving the situation not making it worse.


To math it out: 8 hours a day at $3 USD per hour with 2 weeks vacation is about $15,000 per year.

That's not a lot of money to someone who lives in the United States. But here in 2025 it gets you out of the bottom quintile of earners in China, India, Brazil, Russia, Turkey, Japan, Central America, South America, Africa, South Asia, Southeast Asia, most of Eastern Europe...

For a job that's on demand, and requires, as far as I can tell, decent English skills and an Internet connection, but no real barriers to entry otherwise. It would have been a much stronger deal back in 2012, of course.

I'd be interested to know if the introduction of MTurk as a market competitor pushed entry level clerical wages up in some of these areas. Probably not, because English proficiency in a non English speaking country is a very rare skill not usually borne by people in the bottom 20% of income. But that's probably less true today given the dominant of English language YouTube.


It is interesting how much traction this 9B model is getting which is good.

Still two month earlier 19 European language model with 30B parameters got almost no mention:

https://huggingface.co/TildeAI/TildeOpen-30b

Mind you that is another open model that is begging for fine-tuning (it is not very good out of box).


This 30 Euro jump in Europe was a kick in the pants for me.

Even though it is still a relatively good deal for a Family Plan (compared to say Google Drive or Dropbox) for OneDrive, I finally dropped my Microsoft 365 Family plan.

The final straw was that the Copilot was completely unhelpful and hallucinated features Office portal does not have.


Not OP, but CLIP from OpenAi (2021) seems pretty standard and gives great results at least in English (not so good in rarer languages).

https://opencv.org/blog/clip/

Essentially CLIP lets to encode both text and images in same vector space.

It is really easy and pretty fast too generate embeddings. Took less than hour on Google Colab.

I made a quick and dirty Flask app that lets me query my own collection of pictures and provide most relevant ones via cosine similarity.

You can query pretty much anything on CLIP (metaphors, lightning, object, time, location etc).

From what I understand many photo apps offer CLIP embedding search these days including Immich - https://meichthys.github.io/foss_photo_libraries/

Alternatives could be something like BLIP.


Like Simon I've started to use camera for random ChatGPT research. For one ChatGPT works fantastically at random bird identification (along with pretty much all other features and likely location) - https://xkcd.com/1425/

There is one big failure mode though - ChatGPT hallucinates middle of simple textual OCR tasks!

I will feed ChatGPT a simple computer hardware invoice with 10 items - out comes perfect first few items, then likely but fake middle items (like MSI 4060 16GB instead of Asus 5060 Ti 16GB) and last few items are again correct.

If you start prompting with hints, the model will keep making up other models and manufacturers, it will apologize and come up with incorrect Gigabyte 5070.

I can forgive mistaking 5060 for 5080 - see https://www.theguardian.com/books/booksblog/2014/may/01/scan... . However how can the model completely misread the manufacturers??

This would be trivially fixed by reverting to Tesseract based models like ChatGPT used to do.

PS Just tried it again and 3rd item instead of correct GSKILL it gave Kingston as manufacturer for RAM.

Basically ChatGPT sort of OCRs like a human would, by scanning first then sort of confabulating middle and then getting the footer correct.


Yeah, I've been disappointed in GPT-5 for OCR - Gemini 2.5 is much better on that front: https://simonwillison.net/2025/Aug/29/the-perils-of-vibe-cod...


Images in general, nothing comes close to Gemini 2.5 for understanding scene composition. They perform segmentation and so you can even ask for things like masks of arbitrary things or bounding boxes.


To be a bit flippant, you can absolutely destroy energy by creating some mass..

Then again most of us do not have particle accelerator nearby looking for Higgs boson.


>> The energy you use to flip the flip flop

> To be a bit flippant

I see what you did here :)


I’m sorry, but no. Mass is just energy.


Pretty sure you also need about 120 lines of prompting from default.yaml

https://github.com/SWE-agent/mini-swe-agent/blob/7e125e5dd49...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: