lovasoa

The thing I would have liked to know is why they don't use an existing fast SQL parser. Was being slightly incompatible with all existing SQL dialects a product requirement?

3) The engineers working on the chip used ChatGPT from time to time.

I tried a few PPTX files from consulting firms that are available online. The rendering does not seem to be truly pixel-perfect, but all of them were quite readable and had a good layout, which is already an impressive feat.

The details are interesting: https://en.wikipedia.org/wiki/Eskom#Corruption

How much faster/slower are you with that process compared to writing code yourself?

I use the latest codex with gpt5.4 and Claude opus every day. they hallucinate every day. If you think they don't, you are probably being gaslighted by the models.

I was interested in the human results, so I had an llm build a visualization for them: https://codepen.io/lovasoaaa/pen/QwKWGBd

You can see that 17% of answers come from India alone and that software developers got below average results, for instance.

The author of this project is also the author of redis. He knows what he is doing.

Running inference for a model, even when you have all the weights, is not trivial.

I use linux at home (with a HiDPI screen) and MacOS for work. The screen works well with both computers. I mostly just use a text editor, a browser, and a terminal though.

Linux has bugs, bug MacOS does too. I feel like for a dev like me, the linux setup is more comfortable.

How do you call the opposite of green washing? When you want to show that you are burning as much energy on training models as the others.

Because we can

I use a fork of sqlx in SQLPage [1]. I think my main complaint about it is runtime errors (or worse, values decoded as garbage) when decoding SQL values to the wrong rust type.

* [1] https://sql-page.com/

A consortium of investigative journalists investigated and attributed the attacks the Ukraine's secret service: https://en.wikipedia.org/wiki/Nord_Stream_pipelines_sabotage

No. The LLM's answer is correct.

No it is not: https://en.wikipedia.org/wiki/FIFO_%28computing_and_electron...

In my opinion, the clean way to implement this is with methods instead of attributes for name, desired_speed, etc...

In the second handpicked example they give, GPT-4.5 says that "The Trojan Women Setting Fire to Their Fleet" by the French painter Claude Lorrain is renowned for its luminous depiction of fire. That is a hallucination.

There is no fire at all in the painting, only some smoke.

https://en.wikipedia.org/wiki/The_Trojan_Women_Set_Fire_to_t...

If the electricity were generated by thousands of volunteers pedalling in their basement, then yes, I would expect the utility company not to be too greedy.

I tried passing just the plot to several models:

    ChatGPT (4o):      Noticed "a pattern"
    Le Chat (Mistral): Noticed a "cartoonish figure"  
    DeepSeek (R1):     Completely missed it
    Claude:            Completely missed it
    Gemini 2.0 Flash:  Completely missed it
    Gemini 2.0 Flash Thinking: Noticed "a monkey"

This resonates with me. I used to love tinkering with "underdog" tech. I did frontends in svelte before it was cool. But for new projects, I now always default to react+tailwind, because that's what Claude knows best. And for the backend, I'm working on a SQL-only framework, because LLMs are now at almost 100% accuracy on text-to-sql.

And since the gold supply is limited, fewer shovels would indeed be required.

Law enforcement could probably just ask cloudflare for the exact IP address that retrieved the attachment.

The principle of the software seems to be that the original data is never altered. It is a postgres extension that "masks" the data for certain postgres users. You can always connect as the root user and see everything when you need to.

Does anyone here know more about the unreliability of Google's sub claim? I am using it, and have never had issues with it, but this post seems to just dismiss it without more details.

When does Google change the subject of a token for a given user account?

Dashes in API keys are really annoying. Double-clicking doesn’t select the full key, which just adds extra hassle. It would be much better if they used a continuous string without any separators. Makes copying and pasting way easier, and doesn't affect security at all.

I have never heard of anyone in the developed world who would be afraid to criticize Al-Qaeda or the Taliban or Isis. Actually I have never heard anyone NOT criticize them when talking about them.

Making a good user interface is definitely not easy. Yet it's orders of magnitude easier than writing ffmpeg.

That said, there is nothing wrong with a paid wrapper around a large and complex open source library. Distributing their work more widely is not a disservice.

The best solution, that the LLM did not find, is

     def find_difference(nums):
         try: nums.index(3999), nums.index(99930)
         except ValueError: raise Exception("the numbers are not random")
         return 99930 - 3999

Starting with a Google Sheet is perfectly fine, but for building a non tabular user interface, or having a few thousand records, I would directly switch to something like SQLPage with a SQLite database.

The researchers leverage existing language Chain-of-Thought data, where each sample consists of a question, reasoning steps, and the final answer. At stage 0, the model does not generate any thought tokens, and is just trained to yield the reasoning traces and correct answers for the Chain-of-Thought samples. In the subsequent stages, at each stage, we remove one reasoning step from the sample, and instead add thought tokens. In the illustration above, a single thought token is added in each stage, instead of a single reasoning step, but this is controlled by a hyperparameter ‘c’.