HNHacker News
TopNewBestAskShowJobs

AStrangeMorrow

267 karma · joined June 27, 2016

submissionscomments
AStrangeMorrow··on The Tesla Model Y Just Scored the Worst Reliability Rating in a Decade
Yeah people love to hate on Tesla and always point out the maintenance issues. But from the reports I have seen, Tesla is indeed one of the companies with the worst record in terms of number of incidents requiring maintenance, but among the cheapest, if not the cheapest in cumulative amount spent on maintenance.

I feel like if these report included the price and severity instead of just the frequency of the maintenance it would paint quite a different picture.

AStrangeMorrow··on Claude Skills
I think, from my experience, what they mean is tool use is as good as your model capability to stick to a given answer template/grammar. For example if it does tool calling using a JSON format it needs to stick to that format, not hallucinate extra fields and use the existing fields properly. This has worked for a few years and LLMs are getting better and better but the more tools you have, the more parameters your functions to call can have etc the higher the risk of errors. You also have systems that constrain the whole inference itself, for example with the outlines package, by changing the way tokens are sampled (this way you can force a model to stick to a template/grammar, but that can also degrade results in some other ways)
AStrangeMorrow··on Fei-Fei Li: Spatial intelligence is the next frontier in AI [video]
Yeah even LLM generated code for a 2D optimization problems with many spatial relationships has being absolutely terrible, while I had great success in other domains.
AStrangeMorrow··on Fei-Fei Li: Spatial intelligence is the next frontier in AI [video]
Yeah funnily I did a project where we had an LLM based interface to in house 3D parametric modeling system and it did fairly well.

However I’ve been trying to use LLM, both as orchestrators and in other cases to write code for 2D optimization problems with many spatial relationships and it has done terribly.

I have talking it can generate 1000s of lines over many rounds of prompting/iteration that solve maybe 30% of the problem (and the 30% very easy cases) while completely messing up the rest. When doing that code myself, in less than 1000 lines, the “30% part” was maybe 3% of the total code. Even when basically providing pseudo code to solve specific part of the problem chances are these LLM solutions would also have many blind spots or issues.

The thing is, that is a 2D problem for which there basically no ressources about online, and all the slightly similar problems all have careful handcrafted specialized solutions. So I think it has no good frame of reference how to solve the problem

AStrangeMorrow··on The Metamorphosis of Prime Intellect (1994)
Yeah I really liked that book, but honestly almost dropped it after the first chapter for that very reason. I get the point that is being made with all this gratuitous violence in the context of the whole book, but still
AStrangeMorrow··on Pentagon to terminate $5.1B in IT contracts with Accenture, Deloitte
Yeah sadly I have a similar experience from my school. People going into consultancy were either people that just did software engineering to quickly move into a management position and make more money, or people that were not passionate about the domain or the students that under performed. We’d have consultancy agencies come advertise basically saying “don’t worry we have LOTS of positions open”.
AStrangeMorrow··on Pentagon to terminate $5.1B in IT contracts with Accenture, Deloitte
Yeah. Also might be a bias from my school, but for the software engineering specialty, in average the worse a student was performing the more likely they’d end up working for a consulting company after graduation. A few exceptions (student with a goal of moving into management positions) but none of the students that were either very good or passionate are working in the consulting world now. I am sure there are some skilled consultants, and these companies have their use, but I am always shocked to see how much governments spend on them.
AStrangeMorrow··on Agno: Agent framework 10,000x faster than LangChain
I’d wager it can probably a few percents of the full runtime. But no matter the variation in the time it takes for LLMs to generate outputs (depending on the nb of tokens produced/input size) likely drowns it completely.

Like 5s+/-1s vs 4.95s+/-1s

AStrangeMorrow··on Hallucinations in code are the least dangerous form of LLM mistakes
For me it has helped with: - boilerplate/quick prototyping - unit tests - quick start with a new language/library that I don’t really know

But yes, once the codebase starts to grow ever so slightly the only use I found is a glorified autocomplete.

Overall it does save me time in writing code. But not in debugging.

AStrangeMorrow··on Probably pay attention to tokenizers
Yes, used to work on a system that has elasticsearch and also some custom Word2Vec models. What had the most impact on the quality of the search is ES and on the quality of our W2V model were tokenization and a custom ngrams system.
AStrangeMorrow··on Synthetic diamonds are now purer, more beautiful, and cheaper than mined
I have seen lab grown diamond being quite a bit cheaper than mined ones for a while now. As in ×2 to ×3 times cheaper even.

And yes, funnily it seems that the purer a diamond is (clear, few impurities etc) the higher its price/carat, until it is so pure that it means it is a lab grown diamond and not a natural one and the price drops

AStrangeMorrow··on Structured Outputs in the API
At least depends on the approach and use: stuff like outlines (https://github.com/outlines-dev/outlines) that actually changes the sampling to adhere to a grammar and can be used with local/custom models shouldn't be too impacted. Those are not really used on top of openAI models
AStrangeMorrow··on Structured Outputs in the API
Also the outlines library: https://github.com/outlines-dev/outlines
AStrangeMorrow··on Ask HN: What is the most challenging programming course you have taken?
Back in University the prolog class required us to pick a game and write a small "AI" opponent for this game in prolog.

Either people loved the whole different paradigm of logical programming and were really invested or they just didn't want to touch it. There wasn't much in-between.

I think it is great to expose people to different things and I loved it, but I agree for many people that sent prolog straight to their blacklist

AStrangeMorrow··on It’s an age of marvels
I am just glad to see I am not the only one making up imaginary scenario where I bring people from the past to have them marvel at our world.
AStrangeMorrow··on Liu Cixin's War of the Worlds (2019)
Yeah, I did enjoy the books, but the characters are not great and the prose pretty basic, the romances fall flat etc. But the more you go on the more interesting science fiction are brought up.

Overall I thought that these were some great and thought-provoking science function ideas wrapped up in an average to poor literary package.

AStrangeMorrow··on Wayfair fired a bunch of people again today, after using them to train AI
I think the issue that most people have is also that companies expect a certain level of respect and dedication from their employees but are not willing to reciprocate.

Make sure to work as hard as possible for us, pleas work extra hours if the company is doing bad. Don't take your vacations when it might impact team performance badly. If you leave give up a big heads up.

However we'll refuse your raises and bonuses if we feel like it. Also must be willing to move and uproot your whole life with no guarantee we'll keep you for long. We'll fire you one day after telling you your job was safe, and close all your company accounts before you get to save any of your documents or say bye to your colleagues. And remember no complaining after you leave, or we'll get you. Happened to quite a few people I know, not me thankfully.

And just "get an other job if you company sucks" doesn't always work. You might not know that your company sucks until you're out the door, or finding a new job in your field or without moving your whole family might not be easy

AStrangeMorrow··on Wayfair fired a bunch of people again today, after using them to train AI
As much as I agree that people should try to avoid companies with bad ethics, it is not always that simple. I know some junior in tech that need to send 300+ applications for even just an internship and it is not uncommon for even more seasoned people to have to send 100s of applications to just even get a decent number of interviews. It is just hard to get a perfect view of the ethics of these hundred of companies. And if you just do that check at the interview step, which I am sure quite a few people will do, it can be tough to refuse 80% of your interviews because of companies have bad practices, when you struggled badly to land these few interviews.
AStrangeMorrow··on Plastic bag bans work
Yes. I feel like people often solely focus on CO2 emissions as the only metric. Especially some detractors of ecological pushes and regulations. Like "it doesn't really decrease CO2 output, so what the point", or "it's such a tiny fraction compared to X industry of Y country, and such an inconvenience so what's the point". I just don't understand how producing less stuff whose sole objective is to end up, in the best case, directly in landfill, is bad
AStrangeMorrow··on Plastic bag bans work
I guess that depends on habits and preferences, but I much prefer my sturdy reusable bags over the basic plastic bags.

Sure I do sometimes forget to bring my reusable bags, but even environmental reasons aside I hate plastic bags. It makes it harder to carry groceries, where you might have 10-20 bags instead of two big solid bags. They also frequently rip, so you have to pick stuff up and your groceries gets damaged. Then you have to toss all of them (or store them somewhere "in case" where you end up with 100s of them).

AStrangeMorrow··on Rope making in the Aurignacian of Central Europe more than 35,000 years ago
It is a bias that many people with only a passing interest in history sadly have: that the more far back you go the dumber/less sophisticated people were.

The more you look at historical evidence the more you see that, quite logically, people were very much like us. Complaining about the same stuff ("kids these days...") and so on. Sure a 12th century European farmer would probably struggle with a bit with a smartphone (same as many of our grandparents), but had likely a huge understanding of what plant required what environment to thrive, their growth cycles, what was edible and how to safely prepare it, how to do manual labor while keeping your clothes impeccable etc.

AStrangeMorrow··on Midlibrary: Library of 4000 Midjourney artistic styles
Ethical concerns aside (which is a big topic by itself) having been quite invested in the generative scene for years now, no matter what some people say there is a core of more power users that push to experiment with the technology, the styles, the limitations and the weirdness of it. Not that it has crazy complexity, but some still.

But at the same time the barrier to entry is really low. I now see AI art everywhere: in articles, on LinkedIn, on posters etc. But most of the time in the very most basic style. It feels like instead of searching for an image of X people just type "image of X" in MJ / DallE3 and co, and go with whatever generic stuff they get

AStrangeMorrow··on In Indonesia, humanity's oldest art is flaking away. Can scientists save it?
Reminds me of the Chauvet Cave in France, which also holds cave-painting, where the large number of visitors changed the atmosphere of the cave, causing mold and damaging the paintings.

To preserve it a replica cave ended up being built for tourism, with the original being closed off except for scientific excursions

AStrangeMorrow··on Tons of dead sardines, mackerel wash ashore in northern Japan
Yes, that whole Fukushima radioactive water release polemic was a little bit ridiculous.
AStrangeMorrow··on Tons of dead sardines, mackerel wash ashore in northern Japan
From what I remember wasn't it just drama for the sake of drama and general fearmongering?

The radioactive water that was released was treated and diluted, to have a radioactivity way below what is allowed and safe for ocean release. The release was not exceptional, neither for Fukushima nor for the many other nuclear power plans that do similar water release.

China especially routinely releases water from its nuclear power plant with higher tritium concentration than Fukushima.

Ex of discussion: https://www.reddit.com/r/NuclearPower/comments/169ekpv/gener...

AStrangeMorrow··on Napoleon: Two centuries of life after death
Exactly. All of the battles / wars during the early Napoleon reign were defensive. Out of the 7 coalition wars he only started two.
AStrangeMorrow··on Despite just 5.8% sales, over 38% of bug reports come from the Linux community (2021)
I would disagree. Windows users care about the bugs, but not about reporting them. How many times do I see reviews on steam like "great game but too many bug / game breaking bug X" so 2/10.

Most users just experience the bugs, get mad about the game and move on. Being able to fix the bugs fast and early means less users will experience them in the long run, which in turns means a better gaming experience and very likely better player retention and higher review/player attraction

AStrangeMorrow··on Realtime Planet Shader: Interactive 3D planet animation
Wondering too. It does look like it though. It has that noisyness / fuzziness that makes it sometimes a bit too random compared to earth continents. I found that it works best for earth like map simulations (with water basically) when combined with some terrain physics (moving plaques of random sizes and other geological elements) to generate the high level details, and Perlin on top for the natural noisiness. For other planets though it does great.
AStrangeMorrow··on Oldest Pyramid Found Not in Egypt or Americas, but in Indonesia
True. Just because someone uses some else research to push a bad narrative doesn't mean the initial research is bad.

But in this case it fairly certainly is. Either by lack of scientific rigor, if not worse.

Doesn't make much sense from an archeological standpoint, and there are concerns from the geological standpoint too it seems:

Ex: https://www.reddit.com/r/geology/comments/17t81dg/question_r...

AStrangeMorrow··on Oldest Pyramid Found Not in Egypt or Americas, but in Indonesia
Also this thread on the topic, debating the claim from a geological standpoint' https://www.reddit.com/r/geology/comments/17t81dg/question_r...
← PreviousPage 3 of 5Next →