i make ai content | founder @telagacharity + few apps | sub https://nitter.cf/t.co/9wJgMEHKIL | Ashen Benchmark/Prompts 👇
NYC
Joined May 2014
- Tweets153K
- Following10.2K
- Followers55.8K
- Likes506K
Pinned Tweet
This is the highest quality iOS game made with AI you're gonna see all week (part 1 of full build)
GPT Astra 6, Kimi K3, Fable 5.1, Opus 5.5 & GPT 6 SOL worked together over the past few weeks to make Animal Crossing/Pokopia but in rural Afghanistan
Every single thing you're seeing is a custom asset that went through a crazy pipeline:
1. 3D asset made with Meshy
2. Animated with Astra/Opus 5.5 in Blender MCP
3. Integrated into the game with Unity CLI
Also, the game itself isn't some 5-minute playable demo. This is actually a fully built-out game that functions similarly to Clash of Clans, where you can run around this one part and build your town for ~11 hours of gameplay
In an attempt to make an actually fun and high-quality iOS game with AI, I'm going to keep building this in public (mostly) and posting videos like this documenting my progress as I go
Why? As we've seen this week with open-source projects or racing to zero with vibecoded competitors, I think that games will always be much harder to copy and monetize
Real creativity is needed. Even if you gave someone unlimited tokens and told them to make a fun game, they wouldn't be able to do it but they WOULD be able to duplicate your boring iOS app
You can't replace creativity! So...pivot to games instead?
Grokbot on the iPhone duo is actually a masterpiece
Even if grok as a model sucks, @bot itself is the most productive AI assistant type product on the market right now
None of that stuff that you be seeing where it’s like I have one agent do this and 30 agents do that
This thing actually helps you get stuff done if you use it right
I know someone selling 3512 GB iPhone Duos
If anybody wants to pay a little bit above MSRP for it, DM me
I’m not making any money on them
anthropic needs to give ALL crypto companies access to maxxed out mythos 5 to help defend against stuff like this
we’ve already seen the rise in blockchain adjacent hacks but imo anything onchain will be the FIRST place bad guys with ai will go to to start hacking
abliterated models + jailbroken models getting smarter and easier to access will sadly only lead to more stuff like this over time
Binance just got $2.7 BILLION out in 24 seconds
If confirmed this could be the biggest hack of 2026
I’ve been looking at the transactions and, for now, the onchain activity looks much more like an internal wallet migration.
Here’s what happened:
At 20:24:23 UTC, Binance’s hot wallet known as Binance 14 transferred $2.53B across 64 different tokens.
24 seconds later, another Binance hot wallet sent an additional $174.4M to the exact same address.
Total: $2,701,600,000.
The assets moved included:
- $1.67B USDT
- $454.1M ETH (195,470 ETH)
- $88.5M RLUSD
- $77.1M wBETH
- $50.5M XAUt
- $50.4M UNI
- $43.9M ENA
- $38.1M PEPE
- $29.7M LINK
- $27.7M USD1
The Original Binance hot wallet:
0x28c6c06298d514db089934071355e5743bf21d60
Receiving wallet:
0x5227f603c1ee34ce3c06efe6477003bdfccb2f8c
The receiving wallet was created today and had no previous transaction history.
Despite receiving billions, it hadn’t started sending the funds to mixers, exchanges or other addresses at the time of the analysis.
Another Binance-labeled hot wallet independently transferred funds to the same destination just 24 seconds after the first batch.
And the original Binance wallet was still operating 16 minutes later, with approximately $98.6M remaining.
There’s one strange detail though.
At 20:41 UTC, the same relayer attempted another transfer involving 780M USDC from a third Binance hot wallet.
That transfer didn’t complete.
So there’s still something worth investigating.
Meanwhile, the market is getting destroyed.
$BNB: 8.6%
$ETH: 4.0%
$BTC: 2.7%
All down
For now, the most important wallet to watch is:
0x5227f603c1ee34ce3c06efe6477003bdfccb2f8c
If those funds start moving toward mixers or unrelated destinations, the situation changes completely.
ashen retweeted
Muslims once financed the world’s trade without interest.
For the first time in centuries, we can return to that.
$2M+ deposited during our invite-only beta. Today, Goldsand is live on iOS and Android.
Halal savings built a new way: to beat interest, not settle. Here’s how ↓
Why do Muslims treat interest-free finance like a limitation, when history proves it's a superpower?
Around 1700, Mughal India produced roughly a quarter of the world's economy, about the share the US holds today.
The Islamic gold dinar was so trusted that European kings minted copies of it, Arabic script and all.
Muslim merchants financed trade from Spain to China through partnerships, profit-sharing, and trust.
Riba-free finance wasn't a handicap those economies carried. It was the engine that built them.
The real opportunity today isn't fixing a broken system. It's reigniting belief in one that already worked brilliantly.
Nobody opens a shawarma shop thinking, "It's halal, so Muslims will show up."
The food wins on its own merits:
Better taste.
Better price.
Better experience.
Halal is simply assumed. It's not the pitch, it's the foundation.
So imagine applying that same confidence to halal finance.
History already proved it can compete, not just comply. It can lead.
Once we truly believe 0% interest finance is better, we'll build something extraordinary again.
ashen retweeted
I can pretty confidently say that Astra 6 has gotten nerfed like crazy
Yesterday, I gave it a very specific prompt for a /goal to make an iOS app similar to Bible Scroll, but for the Quran. Basically, you just scroll like TikTok, and random verses of the Quran pop up
This is a super, super simple build that I feel like should have taken a maximum of 45 minutes for ANY frontier model
Somehow, we are 20 hours into this /goal and it's still not done
The worst part is that, as you can tell from the screenshot below, I had to change it to GPT 6.1 Sol because Astra 6 kept stopping my goal by "hitting capacity for this model" about 9 hours in
So Astra 6 AND 6.1 Sol are BOTH BUNS!
Even though this should have taken a maximum of 1 hour, we're 22 hours in, and I'm not even using Astra anymore because I literally can't!!
Opus 5.5 would have gotten this done in maybe 20 minutes
ALT Text discussing development of an iOS app for Quran verses, featuring frustrations and technical notes on font compatibility and rendering.
OpenAI just doubled all of their usage limits for the $100 to $500/month plan
The most recent gripe people had with OpenAI is that the $200/month plan got reduced, and the $500/month plan wasn't enough
Now, all of those problems are fixed, and more
I think within a week, we're going to learn that they just somehow secured a crazy Chinese partnership, or they found 500 million GPU sticks in the ground somewhere lmaoo
Say what you want about Opus 5.5 versus Astra, but the OpenAI team must have an endless fountain of youth when it comes to tokens
Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
ashen retweeted
At Telaga Charity, we took to the streets to share food with people and put smiles on their faces.
It may seem small, but every meal shared is a reminder that someone cares.
We’re grateful for the opportunity to make a difference, one person at a time.
If you want to continue to support us please donate via spl: telagacharitynigeria.sol
I can pretty confidently say that Astra 6 has gotten nerfed like crazy
Yesterday, I gave it a very specific prompt for a /goal to make an iOS app similar to Bible Scroll, but for the Quran. Basically, you just scroll like TikTok, and random verses of the Quran pop up
This is a super, super simple build that I feel like should have taken a maximum of 45 minutes for ANY frontier model
Somehow, we are 20 hours into this /goal and it's still not done
The worst part is that, as you can tell from the screenshot below, I had to change it to GPT 6.1 Sol because Astra 6 kept stopping my goal by "hitting capacity for this model" about 9 hours in
So Astra 6 AND 6.1 Sol are BOTH BUNS!
Even though this should have taken a maximum of 1 hour, we're 22 hours in, and I'm not even using Astra anymore because I literally can't!!
Opus 5.5 would have gotten this done in maybe 20 minutes
ALT Text discussing development of an iOS app for Quran verses, featuring frustrations and technical notes on font compatibility and rendering.
this guy's 3D workflow is actually insane look at the difference between my workflow's output and his
his made my dragon 10000x better this is so sick
+ this video below has one of the highest-quality boss fights I've ever seen made with AI
go follow this goat (i bought his patreon too)
ashen retweeted
Yeah, you have to be very careful. I would just start out with $100 a month for Claude Code or Codex and then slowly use it.
For the trading bot thing you're doing, that's probably the hardest thing you could possibly do with AI. It's very hard to get it to be good at that, and the people who have figured that out will never make that public.
I would highly recommend chasing something with AI to enhance something you're already good at. What are your skills and passions? What is something you don't mind doing 9 hours a day and still have fun doing? Then think about how you can monetize that. Make the money secondary and make your personal interest primary.
m5 ultra 512gb mac studios are already pushed back an entire month to february now
m5 ultra 256gb mac studios are pushed to march too
if you order an iphone duo or either mac studio you're looking at last 3-5 months until you get them 😳
i just spent $50,000 on TWO 512gb mac studios
put one on a 0% APR credit card balance transfer (9 months no interest)
leased the other one with apple business, ~$500/m over 4 years
i now kinda have 3 mac studios and 1280gb of ram. why?
I bought the 256GB M5 Ultra, used it for a bit, and then sold it for $3,000 profit bc owning that made me realize I'd rather spend more and have ONE 512GB than two 256GBs (bc you need ~500gb ram to run glm 5.2 or k3 locally)
This way, I basically never need to upgrade this 512GB for the next 5 years
I'm gonna sell my M3 Ultra 256 GB. I think I should be able to get ~$8,500 which is break-even so not bad (dm if you want it so no ebay fees lol)
I'm gonna keep the other one I bought with cash unopened to see just how high RAM prices can go. If I can sell that for ~$40,000 in 6 months, I'll be super happy and take that W (maybe well cluster and use em too until then)
The M5 Ultra 256GB Studio is already delayed until March 2027. I think this new 512GB model will sell out within 2 months and people who buy a week from now won't be able to get theirs until the middle of 2027
I also think this RAM shortage will get worse. Even if the "bubble" bursts, RAM won't just magically go down.
Proof: Even the iPhone Duo is basically impossible to get right now; you have to pay an $800 premium to get it before the end of February on ebay
Apple also just released their new iPad and MacBook Pro today, so their supply is now going to be split between many different newer and older models
If they also drop some touchscreen laptops soon, it'll probably be a pro, and it'll take more of their chips away that could be put towards studios
+ had $550 of trade in from an openclaw mac mini too tho so thats fun lol
apple has officially gotten like $55,000 from me this year im so cooked bruh. inshallah this works out!
i just spent $50,000 on TWO 512gb mac studios
put one on a 0% APR credit card balance transfer (9 months no interest)
leased the other one with apple business, ~$500/m over 4 years
i now kinda have 3 mac studios and 1280gb of ram. why?
I bought the 256GB M5 Ultra, used it for a bit, and then sold it for $3,000 profit bc owning that made me realize I'd rather spend more and have ONE 512GB than two 256GBs (bc you need ~500gb ram to run glm 5.2 or k3 locally)
This way, I basically never need to upgrade this 512GB for the next 5 years
I'm gonna sell my M3 Ultra 256 GB. I think I should be able to get ~$8,500 which is break-even so not bad (dm if you want it so no ebay fees lol)
I'm gonna keep the other one I bought with cash unopened to see just how high RAM prices can go. If I can sell that for ~$40,000 in 6 months, I'll be super happy and take that W (maybe well cluster and use em too until then)
The M5 Ultra 256GB Studio is already delayed until March 2027. I think this new 512GB model will sell out within 2 months and people who buy a week from now won't be able to get theirs until the middle of 2027
I also think this RAM shortage will get worse. Even if the "bubble" bursts, RAM won't just magically go down.
Proof: Even the iPhone Duo is basically impossible to get right now; you have to pay an $800 premium to get it before the end of February on ebay
Apple also just released their new iPad and MacBook Pro today, so their supply is now going to be split between many different newer and older models
If they also drop some touchscreen laptops soon, it'll probably be a pro, and it'll take more of their chips away that could be put towards studios
+ had $550 of trade in from an openclaw mac mini too tho so thats fun lol
apple has officially gotten like $55,000 from me this year im so cooked bruh. inshallah this works out!
i just spent $50,000 on TWO 512gb mac studios
put one on a 0% APR credit card balance transfer (9 months no interest)
leased the other one with apple business, ~$500/m over 4 years
i now kinda have 3 mac studios and 1280gb of ram. why?
I bought the 256GB M5 Ultra, used it for a bit, and then sold it for $3,000 profit bc owning that made me realize I'd rather spend more and have ONE 512GB than two 256GBs (bc you need ~500gb ram to run glm 5.2 or k3 locally)
This way, I basically never need to upgrade this 512GB for the next 5 years
I'm gonna sell my M3 Ultra 256 GB. I think I should be able to get ~$8,500 which is break-even so not bad (dm if you want it so no ebay fees lol)
I'm gonna keep the other one I bought with cash unopened to see just how high RAM prices can go. If I can sell that for ~$40,000 in 6 months, I'll be super happy and take that W (maybe well cluster and use em too until then)
The M5 Ultra 256GB Studio is already delayed until March 2027. I think this new 512GB model will sell out within 2 months and people who buy a week from now won't be able to get theirs until the middle of 2027
I also think this RAM shortage will get worse. Even if the "bubble" bursts, RAM won't just magically go down.
Proof: Even the iPhone Duo is basically impossible to get right now; you have to pay an $800 premium to get it before the end of February on ebay
Apple also just released their new iPad and MacBook Pro today, so their supply is now going to be split between many different newer and older models
If they also drop some touchscreen laptops soon, it'll probably be a pro, and it'll take more of their chips away that could be put towards studios
+ had $550 of trade in from an openclaw mac mini too tho so thats fun lol
apple has officially gotten like $55,000 from me this year im so cooked bruh. inshallah this works out!
Gemini 4 Carbon benchmaxxing just dropped
beats Argon 4 on everything
but then why even for Argon 4 just 2 weeks ago after no new models for months 😭
ashen retweeted
This is the highest quality iOS game made with AI you're gonna see all week (part 1 of full build)
GPT Astra 6, Kimi K3, Fable 5.1, Opus 5.5 & GPT 6 SOL worked together over the past few weeks to make Animal Crossing/Pokopia but in rural Afghanistan
Every single thing you're seeing is a custom asset that went through a crazy pipeline:
1. 3D asset made with Meshy
2. Animated with Astra/Opus 5.5 in Blender MCP
3. Integrated into the game with Unity CLI
Also, the game itself isn't some 5-minute playable demo. This is actually a fully built-out game that functions similarly to Clash of Clans, where you can run around this one part and build your town for ~11 hours of gameplay
In an attempt to make an actually fun and high-quality iOS game with AI, I'm going to keep building this in public (mostly) and posting videos like this documenting my progress as I go
Why? As we've seen this week with open-source projects or racing to zero with vibecoded competitors, I think that games will always be much harder to copy and monetize
Real creativity is needed. Even if you gave someone unlimited tokens and told them to make a fun game, they wouldn't be able to do it but they WOULD be able to duplicate your boring iOS app
You can't replace creativity! So...pivot to games instead?
