Herring

23 ms·

481.
by Herring 1y ago
I think it's just a symptom of too much stress/work and too little time. If you want people to slow down, you need to back the usual suspects: work-life balance, 4-day work week, affordable housing, unions, high taxes (countering
482.
by Herring 1y ago
Gemini pro + overleaf (learned latex from a long stint in grad school). Cheers mate.
483.
by Herring 1y ago
That doesn't matter. It's like code syntax highlighting -- A good tool benefits everyone. You can pride yourself on doing it the hard way, like reading monochromatic code, but the world has already moved on. FWIW I think there&#x2
484.
by Herring 1y ago
Then in the A/B test you're the 1 fail out of 7 tries, everyone else loved it, and you haven't even looked at it. I can live with those results. Lesson is don't always publicize the LLM help.
485.
by Herring 1y ago
Gemini pro. The flow chart, color-coding, tables etc are just latex. The illustrations are picked off the web. I suspect I might have to hire a professional photographer eventually.
486.
by Herring 1y ago
I'm sympathetic. It's like if code isn't color-coded and properly indented, I'm just not reading it.
487.
by Herring 1y ago
Idk, I'm more optimistic than the author. I'm currently using a LLM to rewrite a fitness book. It takes ~20 pages of rambling text by a professional coach // amateur writer and turns it into a crisp clear 4 pages of late
488.
by Herring 1y ago
If you can do better please go ahead?
489.
by Herring 1y ago
Verification for code is just running it. Maybe "verification" was the wrong word. The model just needs a sense of code X leads to outcome Y for a large number of (high-quality) XY pairs, to learn how to navigate the space better,
490.
by Herring 1y ago
Code and math are similar to chess/go, where verification is (reasonably) easy so you can generate your own high-quality training data. It's not super straightforward, but you should still expect more progress in coming years.
491.
by Herring 1y ago
A lot of people are really bad at change. See: immigration. Short of giving everyone jazz improv lessons at school, there's nothing to be done. To be fair, change is not always good. We still haven't fixed fitness/obesity iss
492.
by Herring 1y ago
If you have to be persuaded not to do evil shit, well democrats (and independents) increasingly don’t even want you. https://news.gallup.com/poll/692150/american-pride-slips-new... One fine day, liberals will wake
493.
by Herring 1y ago
Different perspective: I'm currently using a LLM to rewrite a fitness book. It takes like 20 pages of rambling text by an amateur writer and turns it into a crisp clear 4 pages of latex with informative diagrams, flow charts, color-cod
494.
by Herring 1y ago
I’ve been all around the world. Only in America do people insist on doing evil things and still want to be praised for it. It’s the stupid brand of evil too, like killing millions AND pissing away trillions of dollars in useless Middle Ea
495.
by Herring 1y ago
Way longer than that if you want to get technical: https://en.wikipedia.org/wiki/Divide_and_conquer In a nutshell that's why the South is so poor. They've been falling for this for generations.
496.
by Herring 1y ago
Step 1: Point at immigrants/trans/blacks/etc Step 2: Cut taxes on the rich. <---------- You are here It works every time. Pres. Lyndon B. Johnson said: “If you can convince the lowest white man he's better than the b
497.
by Herring 1y ago
Yeah I noticed the same. I would much rather have sensitive discussions (eg US foreign policy) with Gemini pro rather than with a human. I was recently watching Seinfeld, and apparently in the 90s, picking up a friend from the airport was a
498.
by Herring 1y ago
This species is going extinct. I finally accepted that when my dad died rather than change his lifestyle, despite being warned 10000x. My mom survived a heart attack, saw what happened to my dad, still hasn't changed her lifestyle.
499.
by Herring 1y ago
I like them for refactoring and “explain this massive codebase please”. Basically polishing or investigating things that already work. But I think we should expect the scope of LLM work to improve rapidly in the next few years. https:&#x2F
500.
by Herring 1y ago
Yes this is a powerful argument against discrete packets of information ... delivered to my browser as a series of discrete file packets (HTML, CSS, JS) that are now cached as a local copy on my machine.
501.
by Herring 1y ago
I just keep looking at this chart https://data.worldhappiness.report/chart The US is steadily becoming more and more unhappy. The solutions are fairly basic and fundamental - fix inequality, green spaces, walkable cities, h
502.
by Herring 1y ago
Cars were/are inevitable. But they did massive damage to human fitness, which we still haven't recovered from. I intentionally don't own one, and at least some places in the world are starting to wake up and restrict them and
503.
by Herring 1y ago
> Not innovating, maintaining high prices, and price-gouging are unstable activities for a monopoly Depends how good their moat is. Or how deep their pockets are, because they can often bribe to keep their competitors out (Intel vs AMD).
504.
by Herring 1y ago
Who cares exactly how they did it. Point is they did it and there's zero trust they won't do it again. > Actually it's a good thing that the model can be easily Nazified This is not the flex you think it is.
505.
by Herring 1y ago
There's an inner ToolRetriever which is a LLM call to select the most relevant tools/data/libraries.
506.
by Herring 1y ago
To be fair, helping people like them is deadly to your community. That's what theyre signalling, and I think they might have a point. You can't just give people money (power) because they're poor - look at Latinos voting for
507.
by Herring 1y ago
I'm a bit bearish on SSMs (and hybrid SSM/transformers) because the leading open weight models (DeepSeek, Qwen, Gemma, Llama) are all transformers. There's just no way none of them tried SSMs.
508.
by Herring 1y ago
I think it depends on your niche and model. Gemini pro worked amazing for me in when doing (relatively simple) graph algorithms in python, but completely sucked when I switched to (relatively complicated) latex layouts.
509.
by Herring 1y ago
Well the topic of conversation was AGI. I have a stem phd and I like to think I’m a strong reader, but Gemini pro can analyze/distill pages of text at least as well as I can, and much much faster. It only seems to fail when it doesn’t
510.
by Herring 1y ago
Life is a lot bigger than just economics. See here for example: https://data.worldhappiness.report/chart The US economy has never been richer, but overall happiness just keeps dropping. So they vote for populists. Do you th

More ›