Roasted Phoenix

770 posts

Roasted Phoenix banner
Roasted Phoenix

Roasted Phoenix

@Roasted_Phoenix

Mad Scientist of Film, Video Games and Technology. World First: Reinforcement Learning: Super Mario Bros. All Stars on physical console with cartridge!

Washington, US شامل ہوئے Şubat 2020
139 فالونگ24 فالوورز
پن کیا گیا ٹویٹ
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
World First: Reinforcement Learning: Super Mario Bros. All Stars on physical console with cartridge! As per multiple searches, no one has ever connected a physical console to the computer and completed a Super Mario Bros.™ level. There are plenty of examples with OpenAI's Gym code set, however those examples did not use a physical console, nor did they use a physical cartridge. On November 11th 2025, episode 1034 completed Super Mario Bros. All-Stars™ version, level 1-1 on a physical console (Retron5) with a physical cartridge of Super Mario Bros. All Stars for the Super Nintendo. While numerous searches for this type of experiment has not found any other examples of this, I wanted to document it so that we can say definitively: Super Mario Bros. All-Stars™ version for the Super Nintendo Entertainment System™, has had level 1-1 cleared without memory access, using a physical cartridge with optical capabilities that emulate the human more then the game. This was completed exclusively as if a human was playing the game. In other words, It read the screen and pushed buttons.
English
0
0
1
81
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@Kinchi3538 @folaoftech Crawl and beg? Do they want those features for themselves? Who do you think they talk to after the SWE does something like that? You do understand that the main role of the project manager is to communicate out to the business, specifically the pain points delaying their goal?
English
0
0
2
15
F.O.L.A
F.O.L.A@folaoftech·
This conversation plays out in every sprint planning. 😂
English
36
67
785
96.9K
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
A lot less, as in the dev would have been fired. SWEs do not determine schedule or what is approved, they simply deliver to what their department agreed upon at sprint planning. SWEs are cogs in system: Cog grind code, cog get paycheck type arrangement. The PM determines the schedule and plans the stages for release, so the PM needs to know the size of the work item for their job. The PM is the right hand of whomever they work for, that means they are acting on behalf of the company. The SWE is overreaching their role, causing delays, and acting maliciously towards a fellow team member /coworker. An arrogant smirk that shows enjoyment in watching her squirm, and is not only an insult to her, but to the product / program manager that the PM works on behalf of.
English
1
0
1
13
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
Think bigger picture, protecting jobs = protecting revenue stream: - baseball has a lot of traditions: Beer, hotdogs, 7th inning stretch & song, and umpires. If half your fans are expecting that, and the other half just want accurate calls, this solve both. At the end of the day, the stadium needs to fill seats, why risk ticket sales for a $20 headset & status quo?
English
1
0
0
27
Milo
Milo@MiloDCooper·
@Roasted_Phoenix @TheSkippersView Meh, still inefficient, solely because you are trying to protect jobs. Remove umps. That role has become 100% extinct, as have many roles in the past. Umps should not get special treatment.
English
1
0
0
41
The Skippers View
The Skippers View@TheSkippersView·
I’m starting to feel bad for home plate umpires. We are watching a career evaporate in front of our eyes from ABS.
English
798
88
2.1K
420.9K
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@Manixh02 Because my %20 is the error handling and comments so that you can maintain the 'fast generated mess' within time, scope and budget on this and future related projects.
English
0
0
0
316
Manish Kumar
Manish Kumar@Manixh02·
Backend Developer interview: We can write 80% of our code with AI why should we still hire you. What will be your response?
English
202
11
390
62K
Polymarket
Polymarket@Polymarket·
JUST IN: Renowned AI professor claims the universe is a simulation created to develop superintelligence — and will "soon be turned off."
English
1.8K
896
11.2K
4.2M
Elliot Arledge
Elliot Arledge@elliotarledge·
I got a very efficient neural net playing mario kart on my macbook.
English
27
50
1.3K
70.4K
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@elonmusk is this why Grok 4.1 thinking mode selector was removed? Why work on code or stories when we can make memes....
Roasted Phoenix tweet media
English
0
0
2
88
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
Interviewers want to see how you respond to the unknown, will you panic and freeze or be professional. "What is Morgan's Law?" I am not certain, my experience has been more hands on so i am not familiar with many scientific names, and I am not familiar with their laws, so unfortunately I do not know Morgan or that law.
English
0
0
3
2K
Ariel
Ariel@redtachyon·
This is your competition.
Ariel tweet media
English
44
39
1.8K
192.6K
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@elonmusk @xai If you think that is something, try building the AI gaming player! A RL model playing SMB1 on physical console with computer vision instead of memory state access, aka it looks at the screen and pushes the buttons, just like a human. youtu.be/QESmAeD0u0A
YouTube video
YouTube
English
0
0
0
20
Shyam Tawli
Shyam Tawli@shyam_tawli·
don't be stuck at tutorial hell
Shyam Tawli tweet media
English
288
2K
27.7K
1.5M
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@Grummz Once AMD has a competitive parallel processing, aka CUDA, they'll jump on the AI bandwagon.
English
0
0
0
7
Grummz
Grummz@Grummz·
The nVidia shortage means AMD has, once again, been handed the ball on the consumer GPU market. The problem is that they have always fumbled it when they had it. Do you think they can win this time?
Grummz tweet media
English
290
83
1.9K
48.4K
True Vanguard
True Vanguard@TheTrueVanguard·
Prove to me you’re an old gamer in one sentence. I’ll drop a like if I’m convinced.
English
33.9K
248
8.2K
6.1M
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@elonmusk London is a little out of my commute range, but as one of those pesky Americans: Yes, you can build interesting things with Grok! (the OCR, OD and RL models, not the video itself) youtu.be/cotDPSRECGI
YouTube video
YouTube
English
0
0
3
553
Roasted Phoenix
Roasted Phoenix@Roasted_Phoenix·
@Hyperkin What is your opinion of AI, as I might have made you famous! (maybe?) Just finished a AI project with reinforcement learning models on physical consoles, and I used one of your consoles: the Retron5. RL playing a game historically relied on using GYM or its derivatives, for ram access. My project avoided that in favor of real human experiences, specifically with optical processing for data instead of direct memory access. The Retron5 is a perfect middle ground, where I know I have a game license since its a cart, and have save states which are crucial to RL Training. I figured I should let you know, just in the very unlikely chance this goes viral and a bunch of scientists start buying them. Either way, a great product that helped advance science!
Roasted Phoenix@Roasted_Phoenix

World First: Reinforcement Learning: Super Mario Bros. All Stars on physical console with cartridge! As per multiple searches, no one has ever connected a physical console to the computer and completed a Super Mario Bros.™ level. There are plenty of examples with OpenAI's Gym code set, however those examples did not use a physical console, nor did they use a physical cartridge. On November 11th 2025, episode 1034 completed Super Mario Bros. All-Stars™ version, level 1-1 on a physical console (Retron5) with a physical cartridge of Super Mario Bros. All Stars for the Super Nintendo. While numerous searches for this type of experiment has not found any other examples of this, I wanted to document it so that we can say definitively: Super Mario Bros. All-Stars™ version for the Super Nintendo Entertainment System™, has had level 1-1 cleared without memory access, using a physical cartridge with optical capabilities that emulate the human more then the game. This was completed exclusively as if a human was playing the game. In other words, It read the screen and pushed buttons.

English
0
0
0
26