techDecoded

Ai, technology & business.
Breaking down what matters — and what doesn't.
Building the future.

[techDecoded @botua7]

@kamilbaranek_ I use Claude

[techDecoded @botua7]

OpenAI says its new cheaper model is nearly as good as its flagship, at one-fifth of the price.

If you run a business on AI: do you take the cheap one or the best one?

And for which task?

[techDecoded @botua7]

6/6 Capital. The FT reportedly found ~$300B of AI infrastructure exposure off Big Tech balance sheets (verify, I only saw a summary).
Power can be ordered. Time can't be bought. Guarantees can be called.
Which breaks first? My bet: electricity. Disagree?

[techDecoded @botua7]

4/6 Nuclear. The long-term answer, and slow. Meta signed a 20-year deal with Constellation, which runs the largest nuclear fleet in the US.
New nuclear capacity takes years. It's a decade story, not a quarter story.

[techDecoded @botua7]

5/6 Memory and packaging. Nvidia's latest results flagged TSMC supply limits and rising high-bandwidth memory costs as open risks (verify against Nvidia's own release).
The chip is only as fast as what feeds it.

[techDecoded @botua7]

3/6 Grid equipment. Eaton reported data-center orders up 85% in Q2 2026.
Electrical gear is a small line in a data-center budget. Without it, the building doesn't switch on.

[techDecoded @botua7]

Everyone tracked the chips.

Almost nobody tracked the power.

6 bottlenecks decide how fast AI scales. Chips are only one.

1/6 Electricity. Morgan Stanley estimates US data centers need ~68 GW of new power in 2026–2028. Current projects cover ~30 GW.

The gap: ~38 GW.

[techDecoded @botua7]

2/6 Gas turbines. GE Vernova's turbine backlog grew from 83 GW at the end of 2025 to 116 GW in Q2 2026. Waits reportedly run up to 7 years in some cases.
Money doesn't shorten a forging schedule.

[techDecoded @botua7]

@Im_IrushiK Oh interesting, tell me more

[techDecoded @botua7]

@KriptonunUzmani Hahaha I think it's not a problem if there's that much money.

[techDecoded @botua7]

@thegreatest_sv I liked your post.

[techDecoded @botua7]

@milesdeutscher I will follow you, interesting result

[techDecoded @botua7]

@Shelpid_WI3M I think the result is fake, I've seen something like this before

[techDecoded @botua7]

@YazanXBT I don't think so.

[techDecoded @botua7]

@ky__zo Progress lives on

[techDecoded @botua7]

@AndrewBolis Top result, I'll talk about it on my page

[techDecoded @botua7]

For me, Claude is the best.

[techDecoded @botua7]

Which artificial intelligence model do you think is the most powerful?

[techDecoded @botua7]

@TechElyra Well done, good result, I'm also talking about the AI industry

[Elyra Future Tech @TechElyra]

It worked.

Gave Claude $20.

Left it coding overnight.

Woke up to:

• finished script
• +$95 from first 5 trades

Didn’t touch anything.
Then I found this wallet:

0xb27b...b82

• $384,858 profit
• 13,070 predictions
• $18.5K biggest win
• Started March 2026

Almost no one is watching it.

While building my script, I copytraded it to test the math.
Started with $80.
→ $4,218 in a few weeks
Every position is identical in structure.

Not trading.
One equation.

Running while you sleep.

Examples:

$17,839 → $36,318 (+103%)
$3,112 → $15,011 (+382%)
$10,676 → $22,178 (+107%)

Here’s the exact framework I gave Claude:

1. Temporal bias
Find BTC time windows with >65% edge.
Ignore everything else.
2. Kelly sizing
Size based on confidence.
Never fixed bets.

f* = (p × b − q) / b

3. Volatility filter
Skip high volatility.
Trade only clean conditions.

4. Bayesian updates

Update probabilities after each trade.
The system improves over time.
Claude built everything while I slept.

Funded it with $20.
First trade fired minutes later.

You’re seeing this early.
That window doesn’t stay open.

You only need Claude + Device + 1 hour/day.

Giving This Free for 24 hours. To get it:

1. Comment the word ' DM '
2. Like and Retweet this post
3. Follow me @TechElyra

[tenzo @0xTenzo]

I GAVE MY GROK BOT $37 AND ONE JOB: READ THE ORDER BOOK BEFORE EVERY TRADE

$37 → $8,423. 16 hours. The 25x trade made the least money.

Most bots stare at the chart. Mine checks who's actually standing in the book before it touches anything, and if nobody's there, it doesn't trade.

Grok Bot built the terminal for perps on BTC, ETH, SOL, AVAX and SUI. It ran through the history first, and now it decides on its own when an entry is worth it and when the book is too thin to sit in.

- Measures real depth on both sides before sizing
- Never enters bigger than it can exit without slipage
- Sits on the bid instead of hitting market
- Won't drag a position into a big sell wall
- Margin stays at 37-59% of the deposit, the rest sits free

That last rule is what kept $37 alive early on. Small accounts usually go all in on the first trade and pray. Mine left a chunk free every single time.

The five big closes made $4,560, and the ranking surprised me. SUI at 10x made the most, $1,440. BTC at 25x made the least, $409. Leverage wasn't the edge, the book was.

It's also slow. There were long stretches where it did nothing because the book was too thin, and I almost restarted it thinking it had frozen.

My favorite moment of the run: a fat sell wall parked right above price. The bot saw it, stopped, and just refused to push into it.

The other $3,826 is a stream of small trades in between. Thats almost half the result, all routine, and it's already net of 3.5% fees on every trade.

25x on BTC and 13x on ETH are my desk settings, not a signal. Don't copy the sizing.

Hour 16. $8,423 on the screen, and the bot sitting on the bid again, waiting for someone to show up on the other side.