Pinned Tweet
Created the Mario theme melody using nothing but sine waves and a frequency equation.
Here's how i did it.
• Took the Mario melody’s notes and rhythm.
• Converted each note into a frequency: f(n) = 440 × 2^((n−69)/12).
• Added sine waves at odd harmonics to create the chiptune tone.
• Gave each note a short fade-in and fade-out to reduce clicks.
• Scheduled the notes and rests to recreate the melody.
• Generated 48,000 audio samples per second ie. the audio was sampled at fs = 48KHz.
• Slowed the rhythm without changing pitch just for fun.
The equations synthesized the sound.
and viola! we have Super Mario bros. theme music
6
28
1,908
Not to brag, but I have 3 Codex resets.
imma drown in tokens this weekend.
Can you flex harder?
6
12
344
GPT-4 once charged $60 per million output tokens. GPT-6.1 Sol charges $10.
API prices(chronological) in USD per 1M tokens.
Input / Output:
GPT-1 → No hosted API rate
GPT-2 → No hosted API rate
GPT-3 Davinci → $20 / $20 (historical)
GPT-3.5 Turbo 0613 → $1.50 / $2 (historical)
GPT-4 0314 → $30 / $60 (historical)
GPT-4o → $2.50 / $10
GPT-5 → $1.25 / $10
GPT-6 Sol → $2 / $10
GPT-6.1 Sol → $2 / $10
Other prices are current Standard short-context rates. GPT-3 used one shared rate for prompt and completion tokens.
One car. Same angle. A visual journey through GPT’s evolution.
Artistic visualisation, not a model benchmark.
Took a break, post diwali to touch some water, sand and grass.
Hope you all had a great diwali vacation!!
4
17
401
I had this idea of converting sketch to 3d objects , back when i didnt even know coding
Also a very similar game i used to play called roller coaster tycoon 2 , where you could create and run an amusement park.
A little tribute to both a great game and a good idea ig haha.
sketch to a roller coaster ride, if you would want to try it out just ask me i would publish the link and share with you people.
5
8
313
This is very frustrating for Codex/Claude-code users 🚨
Got a new laptop, signed into Codex with the same account, and expected my projects and chat history to be there but of course they were not present .
Instead, I’m figuring out remote connections and handoffs.
Moving devices should be simple: pick my three projects, bring everything over, keep working.
but modularity is not a choice for harnesses it seems.
I should create an extension which can on a single press do this or if somebody has already figured it out do share?
Day 03 : Unslopping my brain cells
Some good implementation based questions today.
one for ref :
given a matrix :
0 0 0 0
0 1 0 0
1 0 0 1
find area of smallest sub matrix such that outside of it only 0's are acceptable.
2
9
219
Completed 500 verified followers.
Need to work on my content haha, all suggestions are appreciated.
Maybe i should go live on topics related to voice AI ?
If you're in tech, let's connect! 🤝
I'm mostly interested in:
-AI & Machine Learning
-Audio AI & Speech Tech
-Software Engineering & Backend Systems
-DSA & Competitive Programming 🚀
-Computer Networks.
I work as an Audio ML Engineer.
Am currently working on developing Voice agents and
replicating tts models from scratch.
Other than this i spend most of time on my judge as a service i am building to host my own Dsa and competitive programming contest on.
Lets talk about what you are building, always curious to know :)
Everybody needs a platform to unslop.
5k users already.
We now have DSA sheet on Trentorch.com
- No you don't have to pay for anything
- 450+ questions
- 20+ Topics
- Multiple Languages: C++, Py, Java, JS, Go, Rust
- Theory of all topics
Learn and practice across Arrays, Strings, Linked Lists, Trees, Graphs, Dynamic Programming, Greedy, Tries, and much more.
-> We don't charge our users anything.
-> It's completely Opensource
-> No Paywall
1
10
233
GPT 6.1 sol extra high vs 6.1 Terra Extra HIgh
Sol got one iteration. Terra got five.
Same V8 engine task. Same Codex harness.
Both on Extra High reasoning.
MODEL TIME · EST. COST · TOTAL TOKENS
• GPT-6.1 Sol: 18m 51s · $1.21 · 5.66M
• GPT-6.1 Terra: 13m 05s · $1.01 · 10.13M
Terra used:
↓ 17% less estimated cost
↓ 31% less model time
↑ 79% more processed tokens
CHECKS
✓ All five Terra versions preserved
✓ Rod-length checks: both passed
✓ Matched front, top and side views
✓ Dismantling and reassembly shown
✓ JavaScript exceptions during capture: 0
VISUAL DETAIL — my assessment
Sol: detailed original build
Terra: schematic → substantially more detailed
Top: Terra. Bottom: Sol.
Final result first, then every iteration.
Tokens include cached context. Costs are model-only API estimates; reviewer effort, waiting and video production are excluded.
Did the feedback close the gap?
Same V8 engine prompt. Three AI models same harness.
GPT-6.1 Sol vs GPT-6 Astra vs GPT-6.1 Terra.
All on High reasoning.
BUILD TIME · EST. COST · BUILD SIZE
• Sol: 18m 51s · $1.21 · 0.59 MB
• Astra: 29m 23s · $5.51 · 0.57 MB
• Terra: 4m 47s · $0.28 · 2.06 MB
FUNCTIONAL CHECKS
✓ Full 30-second sequence: 3/3 completed
✓ Rod-length checks: 3/3 passed
✓ Pause/resume: 3/3 passed
✓ Follow-up prompts needed: 0
✓ JavaScript exceptions during capture: 0
VISUAL DETAIL — my assessment
Sol: detailed
Astra: detailed
Terra: schematic
Top: Sol. Middle: Astra. Bottom: Terra.
Costs are API-equivalent estimates. Build sizes are uncompressed.
Which one would you ship?
10
14
453
Unslopping my brain day 02😆
Getting close to my next star.
With Ai doing 90 percent of your office work ig its important to remove the webs out.
I have started to like codechef interface more than codeforces.
3
84
Same V8 engine prompt. Three AI models same harness.
GPT-6.1 Sol vs GPT-6 Astra vs GPT-6.1 Terra.
All on High reasoning.
BUILD TIME · EST. COST · BUILD SIZE
• Sol: 18m 51s · $1.21 · 0.59 MB
• Astra: 29m 23s · $5.51 · 0.57 MB
• Terra: 4m 47s · $0.28 · 2.06 MB
FUNCTIONAL CHECKS
✓ Full 30-second sequence: 3/3 completed
✓ Rod-length checks: 3/3 passed
✓ Pause/resume: 3/3 passed
✓ Follow-up prompts needed: 0
✓ JavaScript exceptions during capture: 0
VISUAL DETAIL — my assessment
Sol: detailed
Astra: detailed
Terra: schematic
Top: Sol. Middle: Astra. Bottom: Terra.
Costs are API-equivalent estimates. Build sizes are uncompressed.
Which one would you ship?
I prompted GPT-6.1 Sol, GPT-6 Astra and GPT-6.1 Terra to build a 3D rocket launch and landing in the browser.
Model Comparison:
All three used High reasoning. All three reached touchdown.
• Sol: 11m 10s · $1.32
• Astra: 12m 05s · $5.83
• Terra: 9m 30s · $0.28
Terra finished fastest and cheapest, but delivered simpler visuals and less extensive testing.
Costs are API equivalent estimates as i used codex and not my Codex bill.
Top: Sol. Middle: Astra. Bottom: Terra.
Which one actually looks worth the extra time and money?
This is the most value for money laptop accessory purchase yet.
Has a Rj45
Sd card slot
Micro sd card slot
Usb 3.0 port
Usb 2.0 port
HDMI port
2X type C ports and can connect to type c devices.
3
4
226
Amazon Basics 8-in-1 USB-C Hub
⚡ USB 3.0 Data Transfer: 5 Gbps
Theoretical Read Speed: 625 MB/s
Theoretical Write Speed: 625 MB/s
Practical Transfer Speed: ~350–450 MB/s
💾 SD/microSD Read & Write: Not specified
⚠️ Actual speeds depend on your connected device. The hub's USB 3.0 interface is limited to 5 Gbps.
63









