PeteyFoReally

213 posts

PeteyFoReally banner
PeteyFoReally

PeteyFoReally

@petey_fo_really

Used to do computer engineering now I make computer games. Shitty blog: https://t.co/xUAUzQp6wg

Vancouver Katılım Şubat 2014
220 Takip Edilen40 Takipçiler
PeteyFoReally
PeteyFoReally@petey_fo_really·
@WindowsLatest well i guess its better than the literal 1.5 seconds it takes to have windows explorer render. but holy fuck how bad are we at when this is considered good?
English
0
0
1
63
Windows Latest
Windows Latest@WindowsLatest·
Microsoft confirms Windows Run dialog is getting a modern look on Windows 11! - Modern design: A refreshed look that matches Fluent Design and Windows 11, with dark mode support. - Faster than before: Perf was top-of-mind when rewriting Run, and with a **94ms median time-to-show time it’s faster than ever before. - Quick user directory access: You can now type ~\ to jump to your user directory, then keep navigating just like you would from the command line. The modern Run dialog is slowly rolling out in current Insider builds of Windows as an opt-in feature, ensuring we are collecting your feedback. To enable the new Run dialog, you’ll need to: - Be on Windows Insider Experimental Channel. - Enable to Settings -> System -> Advanced and toggle on the new experience with the “Run Dialog” option at the top of the screen.
Windows Latest tweet media
English
108
31
797
497.2K
PeteyFoReally
PeteyFoReally@petey_fo_really·
bbc.com/news/articles/… nytimes.com/2025/07/30/bus… Should probably do your own research instead of taking it from a twitter comment thread, but some of the data collection methodology as a result of significant manpower reductions at the BLS are rather questionable. CPI for example is something like 4% officially but I don't think there is anything a consumer can buy that has /only/ gone up by 4% in the past year
English
0
0
0
17
zöda (in sf!)
zöda (in sf!)@zodattack·
@petey_fo_really when you say its not reliable at this time, are you referencing how the government shutdowns affected the data or something else?
English
1
0
0
21
zöda (in sf!)
zöda (in sf!)@zodattack·
every software engineer i see is nervous about the job market. students are switching majors. twitter makes you feel like you should quit your job and start a company. I dug into the data to see the truth for myself. it's not as bad as you think. it's not great either. excited to share my findings & hear what y'all think:) link below.
zöda (in sf!) tweet media
English
16
5
160
29.4K
PeteyFoReally
PeteyFoReally@petey_fo_really·
@samhogan you're allowed to do both. slop out some code to try out ideas. hand write the stuff that matters.
GIF
English
0
0
15
1.5K
Sam Hogan 🇺🇸
Sam Hogan 🇺🇸@samhogan·
All the best programmers I know are starting to write code by hand again
English
674
344
6.8K
1.5M
PeteyFoReally
PeteyFoReally@petey_fo_really·
@bmcnett do you really? theres about 6000 games on steam released every year with a dev time of about 10 months. are you buying any of those?
English
0
0
0
2
PeteyFoReally
PeteyFoReally@petey_fo_really·
Yeah that's been my experience too. but I think what this whole experience has shown me is surprisingly how many things can get by with just total dogshit in terms of code. even things where I really want something solid, i find using the LLMs to shake out that terrible first draft immediately shows me a lot of issues with what my original plan was. (or validates it)
English
0
0
0
98
Rhys
Rhys@RhysSullivan·
from my experience, even the best models (Opus 4.6, 5.4 xhigh / 5.3 codex) cannot write good code today without an amount of work that is equivalent to just doing the work myself am excited for a world where they can, but in the current state i have very low trust in them
English
277
87
2.8K
321.7K
PeteyFoReally
PeteyFoReally@petey_fo_really·
> hat does a coder who relies heavily on AI tools do in cases like that? Fall back to using their normal coding and debugging skills? Tbh a lot of the vibecoding bros are just slopping it out with javascript frontends, dont even know how to attach a debugger. but i suppose they don't really need it either. What was your problem that you had to debug anyway?
English
0
0
0
72
Vjekoslav Krajačić
Vjekoslav Krajačić@vkrajacic·
So I just debugged a problem that took me a while to figure out. It required inspecting the code through various phases in the debugger. And it made me wonder: what does a coder who relies heavily on AI tools do in cases like that? After AI has given them wrong results after countless "fix this" queries? At what point do you accept that AI won't solve it for you? What if you don't know or don't care what a debugger is? Honest questions btw.
English
102
1
232
27.9K
PeteyFoReally
PeteyFoReally@petey_fo_really·
For human code- If i'm reviewing something im unfamiliar with all i can really do is provide superficial spot checks- ensure that you do your due dilligence and try to ensure your changes are internally consistent. I make this very clear when someone asks me to review their PR that im unfamiliar with- that all i can do is spot checks. if i have time ill ask them to walk me through the code change. AI code is a different beast entirely, it likely wont have any obvious issues i can catch. But rather issues tend to be more regarding alignment and architecture and hallucinations I also notice AI code has insane rte of code churn. I've noticed that its almost pointless to try to keep up b/c its likely going to change in a week anyway. In that case what's important is that the person whos doing the feature is one who's responsible for it and is on the hook for any breakages and has a good ability to keep up with what its needs are.
English
0
0
1
98
Łukasz | Wookash Podcast
Łukasz | Wookash Podcast@wookash_podcast·
To people who are *good* at reviewing code (or claim to be hehe) - how is that possible? To what extend you can properly review the code with low familiarity with the codebase? Eg. New project, you jump in, Claude Code PR - 500 lines changed - review now What's the strategy?
English
178
8
372
97K
PeteyFoReally
PeteyFoReally@petey_fo_really·
Had a strange thought; Code tends to self-replicate, you can write docs on how to implement things, edge cases gotchas etc... and in the heat of the moment. if I really want to do something- I will just look at some code in the codebase that does that already. If the code that's already there is bad, then the coder that references it will just copy that pattern. Doesnt matter if the coder is LLM or human you get the same results. In this way its a good idea to get rid of or tag bad code whenever you see it in your codebases
English
0
0
0
5
Bitcoin Teddy
Bitcoin Teddy@Bitcoin_Teddy·
For all the morons who say capitalism doesn’t work
English
348
668
5.6K
1.3M
PeteyFoReally
PeteyFoReally@petey_fo_really·
@cryptopunk7213 @AnushkaaTyagii The internet's big and twitter only allows for short form ragebait discussion. Most artists i know especially pros have very complex feelings and relationship with AI.
English
0
0
0
68
Ejaaz
Ejaaz@cryptopunk7213·
@AnushkaaTyagii yeah i get it i just don’t agree with the knee-jerk reaction to hate on the technology itself. software engineering is a good comparison, ai is replacing their art but they’re levelling up to orchestrate agents and tools. a lot of them are happier
English
30
0
35
18K
Ejaaz
Ejaaz@cryptopunk7213·
genuine question: why are the arts and gaming communities so fucking touchy about ai? i get that 99% of examples are slop but we’re reaching a point where ai-generated media is indistinguishable for 90% of the world that shouldn’t go un-acknowledged just because you want to hide behind a professional identity ai isn’t going away, it’s getting (a lot) better, so why not try and figure out how to work with it to your advantage? isn’t that what a bunch of hollywood is realising now? doomer: “NO THIS IS NOWHERE CLOSE TO PIXAR GRADE” 8 year old (target audience): “haha that’s awesome” what am i missing?
Is this a 3D model?@IsThisA3DModel

no and this is nowhere close to "Pixar-grade"

English
574
31
977
368.7K
PeteyFoReally
PeteyFoReally@petey_fo_really·
@_trish_xD malloc is just a function with a lot of convention. it is not some holy thing that is enforced by the mechanics of the language itself. almost all serious C projects libraries wrap it anyway. (and a lot of platforms dont even give you a malloc)
English
0
0
0
229
trish
trish@TrisH0x2A·
malloc(0) is legal C. let that sink in for a second. some compilers return NULL. some return a valid pointer you can't dereference but CAN free(). both behaviors are correct according to the C standard. you can allocate zero bytes of memory, get a pointer to nothing, and then dutifully free that nothing. and the language just shrugs and says "yeah that's fine." this is why C developers have trust issues.
English
91
18
640
84.8K
PeteyFoReally
PeteyFoReally@petey_fo_really·
@IamAroke well its generally a bad time to be in games at this moment. epic just laid off 1000 people. outside of established AAA- theres millions of indies releasing close to 30k games on steam per year
English
2
0
0
242
Austin
Austin@IamAroke·
Why are game developers very few? Is game development that very difficult? 🤔
English
52
0
69
14.1K
PeteyFoReally
PeteyFoReally@petey_fo_really·
Its actually really simple. the bottleneck is now your understanding. so instead of using it to do 50 things at once. use it to do the same thing multiple times and go really deep on one topic. Really enjoy working with it this way. also just go take a break when you're done. World was doing fine before this stuff made you 50x faster
English
0
0
0
139
Zara Zhang
Zara Zhang@zarazhangrui·
My single biggest pain point right now: AI-induced attention deficit I have 5 Claude Code sessions running 10 Terminal tabs open 50 browser tabs open 100 X articles in my bookmarks When there are multiple AIs working for you, you're constantly task switching When you're waiting for AI to output, you're constantly distracted Because what used to be hard is now so easy, you always wanna do more, more, more Because of the speed of the AI news cycle, you feel a constant need to be on this app Add all this up and the result: you live in a constant state of distraction & delirium. Focus & deep work is harder than ever If anyone has figured this out pls teach me how
English
517
131
2.1K
240K
PeteyFoReally
PeteyFoReally@petey_fo_really·
@MHBrasil I cant tell if this is a minecraft mod or a monsterhunter mod
English
0
0
1
1.2K
PeteyFoReally
PeteyFoReally@petey_fo_really·
@Hidekazu__ @theo The low level os apis aren't sexy or ergonomic but llms kinda don't give a shit about that it seems.
English
0
0
0
459
Hidekazu
Hidekazu@Hidekazu__·
It’s not hard, it’s that no one gets taught low level languages enough, teaching render APIs and no companies are hiring for those fields as much. If they do create native desktop apps, optimizing and supporting multiple platforms and compilers becomes increasingly difficult to maintain.
English
1
0
4
581
PeteyFoReally
PeteyFoReally@petey_fo_really·
@theo ok but grok's summary of this situation is funny as fuck unless im misunderstanding here.
PeteyFoReally tweet media
English
0
0
5
574
Theo - t3.gg
Theo - t3.gg@theo·
“Everything Claude Code can do, for free” Local model people have lost touch with reality
thestreamingdev()@thestreamingdev

→ Search the web for live sports scores and stock prices → Find files on my desktop and run shell commands → Write code and solve math problems → Everything Claude Code does — for free The breakthroughs that made this possible: Apple's "LLM in a Flash" paper showed models can page from SSD using unified memory. I proved it works in practice on consumer hardware — not just in a research lab. Google's TurboQuant research showed you can compress KV cache with zero quality loss. I applied this with two server flags and doubled my context window from 32K to 64K tokens. For free. No code changes. The biggest surprise: the 35B model at 2.6 bits per weight was supposed to have "broken" tool calling. Every agent framework I tried failed — infinite loops, no answers. I stopped asking the model to generate JSON function calls. Instead I ask it simple questions. "Is this a search, shell, or chat?" → one word answer. Works perfectly. The tool calling wasn't broken. The protocol was wrong. Both models. Full agent. Same $600 computer: → 35B MoE: 30 tok/s, 2x faster, smarter reasoning → 9B dense: 16 tok/s, 64K context, reads entire codebases I benchmarked everything: → 212 math problems: 86.3% accuracy (3 categories at 100%) → 10 web search categories: 10/10 accurate → Shell commands: finds videos, checks disk space, reads code → MLX vs llama.cpp: tested both, llama.cpp wins for 35B

English
114
25
1.3K
181.9K
trish
trish@TrisH0x2A·
first time you wrote hello world - what language did you use?
English
2.3K
210
3.2K
290.8K
PeteyFoReally
PeteyFoReally@petey_fo_really·
1800 lines of c. no stdlib. this produces a binary that is 59KB and it includes a custom sdf based font format that encodes roboto for 117KB, 15MB working set charge when running (seems like its mostly drivers). The entire package together to ship is under 200K and is DPI aware, no dependcies other than what the os provides you already. This is more than sufficient to begin creating a ui system. Possibility here for a very tiny desktop application framework? github.com/peterino2/Arca… #coding #programming #buildinpublic
PeteyFoReally tweet mediaPeteyFoReally tweet mediaPeteyFoReally tweet media
English
0
0
0
61