gymbeaux

5 ms·

31.
by gymbeaux 4mo ago
The PS5 base has 2 more physical cores (4 more threads) than this Steam Machine so I'm not optimistic that the Steam Machine outperforms the PS5. (both "semicustom AMD" so probably effectively the same architecture)
32.
by gymbeaux 4mo ago
Circa 2013 you could get a Steam Machine (made by Alienware/Dell) - the Alienware Alpha - for something like $300. Granted they were clearing them out at that time, but $300 was a no-brainer when consoles cost about the same and had si
33.
by gymbeaux 4mo ago
I recognize that I seem to be an above-average developer and so my standards for LLM output are higher. I'm seeing enough comments where people are comparing LLM output to offshore dev output or even onshore mid-level or junior-level o
34.
by gymbeaux 4mo ago
What I'm hearing is "thoroughly reviewing AI generated code would defeat the purpose, so we give it a cursory glance and it seems to be decent code", and that's my point - it does indeed seem to be decent code but I th
35.
by gymbeaux 4mo ago
True. I used to get code that wouldn't compile all the time from Infosys "developers" in India circa 2016. Perhaps now with LLMs they still do basically no work, but at least the code compiles? That being said, I'm not s
36.
by gymbeaux 4mo ago
If I have to verify the code then I don't see a point in using it to write more than a single method at a time, and that method should be simple enough that I can take a very quick glance and be able to tell it's correct - somethi
37.
by gymbeaux 4mo ago
Yes, thank you for wording it better. When I read through for example an entire codebase that was ~99% written by AI, it's "inconsistent" in a way that even a shared-by-humans codebase would not be. I think this arises from t
38.
by gymbeaux 4mo ago
Those numbers are abysmal. Should we really be using LLMs to write our code? I have a theory- LLMs can spit out code that gets the job done and looks ok, maybe even great, but contains small “anomalies” that compound over time. An enterpris
39.
by gymbeaux 4mo ago
Sure but there’s no way Enron was the last Enron. Also keep in mind something can be “legal” and still misleading to shareholders. I’m sure there are loopholes, some of which may have been reintroduced by Republicans in the last 20 years.
40.
by gymbeaux 4mo ago
(P/E ratios aren’t expressed in dollars) So a PE of 500 means it would take 500 years for the earnings of the company to equal the current market cap (price per share X number of shares). This implies absurd (almost certainly impossibl
41.
by gymbeaux 4mo ago
I think most people just vote for whoever’s mailer they like more. Where I live, the Republican candidate just sends out a mailer with him standing next to a firefighter and a cop and that’s really it. Maybe some garb about reducing crime a
42.
by gymbeaux 4mo ago
This brings up a good point- can we just have Claude Code create a ripoff of this site and run it locally for free? Probably.
43.
by gymbeaux 4mo ago
I think it's worth noting that the "work" these LLMs are doing is digital. ChatGPT cannot plow fields, or construct buildings and roads, or otherwise interact with the physical world. LLMs are useful because humanity has evol
44.
by gymbeaux 4mo ago
Anecdotally I am seeing websites/web portals/et al. are much buggier. I have no doubt that software quality has declined since ChatGPT was made public, and tools like Claude Code are accelerating that "enshittification"
45.
by gymbeaux 4mo ago
I think we should investigate the backgrounds of those making claims one way or another and rely on those backgrounds for determining credibility. I suspect that we'd find that those who are saying LLMs write great, bulletproof code wi
46.
by gymbeaux 4mo ago
The unit test example has been my team's experience as well. The unit tests look good on the surface, but their passing or failing has little predictive value on whether there are actually bugs in the code. Some people have suggested y
47.
by gymbeaux 4mo ago
I’ve been a software engineer for over a decade. At what point is my “residency” over? At what point does the “community” decide I can write software with some level of competency and don’t need to solve fizz buzz over and over again?
48.
by gymbeaux 4mo ago
I think of housing as a better example- KB Home cookie cutter special for $400k or 100-year-old craftsman house for $800k. Most are okay with the $400k cookie cutter special that probably has piss in the walls, because $800k is such a huge
49.
by gymbeaux 4mo ago
Even though the synthetic benchmarks paint a picture of LLMs coming a long way since 2022, my practical experience has been that they aren’t tangibly better. No doubt someone reading this will chime in and say LLMs are way better at writing
50.
by gymbeaux 4mo ago
LLMs seem to be best at writing web apps intended to be internal tools. I don’t need to ever really read the code because the functionality is relatively simple- but still valuable. I had Claude Code build me a CLI tool for running common k
51.
by gymbeaux 4mo ago
LLMs have taken much of the enjoyment out of coding for me, but I don’t think it has to be that way. I hope that as an industry we settle on LLMs being more like tools than human assistants. I think most of the engineers at my company are u
52.
by gymbeaux 5mo ago
Enlighten us to these drawbacks. On the surface I am inclined to say the pros would outweigh the cons. Compared to other professions, software engineering seems to struggle the most with H-1B/Green Card abuse and interview processes. J
53.
by gymbeaux 5mo ago
Great message from the Pope (I never thought I'd write those words). I'm glad he recognizes what's going on with AI (LLMs, really) as one of the highest areas of concern for humanity and the Earth and is using his influence t
54.
by gymbeaux 5mo ago
Surely they’re using a third party mapping library and they have absolutely 0 control over what the borders look like in disputed areas.
55.
by gymbeaux 7mo ago
Strong disagree on AI implying it doesn’t actually work. Staff engineers at FAANG are vibe coding these days and their stuff absolutely works.
56.
by gymbeaux 7mo ago
Fascinating. It’s called Instant Bloomberg and it’s just AOL instant messenger but only between Bloomberg Terminals. So that is absolutely the “data that insiders rely on that pedestrians will never have” and I’m sure it offers tremendous a
57.
by gymbeaux 7mo ago
As far as I know, finding quality data sources is virtually impossible unless you already have so much money that your involvement in the stock market is either because you have insider knowledge or severe boredom owing to your never needin
58.
by gymbeaux 7mo ago
Surely it gets noisy after 30 or so, but pedestrian stock trading apps like Thinkorswim have very high levels of customizability and modularity. I think even extensibility in Thinkorswim’s case. Java I think. Anyway, I would think any users
59.
by gymbeaux 7mo ago
This is cool, but my experience with trading stocks is that you need to spend money in order to be on a remotely-level playing field with the whales. Ignoring that so much stock market data costs money, there’s also the matter of direct fib
60.
by gymbeaux 7mo ago
Which services have 25% uptime?

More ›