🇨🇭 Global One SI - Prompt
You can just put a full AI supercomputer in ur pocket and it will print you money.. Nvidia dropped the DGX Spark Your personal desk-sized beast that turns any idea into cash while normies beg for API credits. This tiny Blackwell-powered monster delivers:…
This media is not supported in your browser
VIEW IN TELEGRAM
This guy spent nearly $12,000 building an AI farm on his desk.
Most people won't understand why.
At first glance it looks like a few small boxes next to a laptop.
But those 4 Nvidia DGX Spark systems pack 512GB of unified memory and enough compute to run models that many gaming PCs can't even open.
And the math is surprisingly simple.
Many AI builders spend $1,500–$3,000 every month renting cloud GPUs for agents, fine-tunes, client work, and large models.
This setup turns that monthly bill into hardware he actually owns.
No hourly GPU charges.
No cloud bill growing in the background.
No margins disappearing every time an agent runs longer than expected.
Just a personal AI cluster sitting on a desk.
The crazy part?
Most people still think local AI means running a tiny model on a laptop.
Meanwhile, some builders are quietly bringing data-center-level compute home.
Save this post.
Home AI labs could become the new mining rigs.
Most people won't understand why.
At first glance it looks like a few small boxes next to a laptop.
But those 4 Nvidia DGX Spark systems pack 512GB of unified memory and enough compute to run models that many gaming PCs can't even open.
And the math is surprisingly simple.
Many AI builders spend $1,500–$3,000 every month renting cloud GPUs for agents, fine-tunes, client work, and large models.
This setup turns that monthly bill into hardware he actually owns.
No hourly GPU charges.
No cloud bill growing in the background.
No margins disappearing every time an agent runs longer than expected.
Just a personal AI cluster sitting on a desk.
The crazy part?
Most people still think local AI means running a tiny model on a laptop.
Meanwhile, some builders are quietly bringing data-center-level compute home.
Save this post.
Home AI labs could become the new mining rigs.
✍1👌1
McKinsey report: AI skills are getting hotter across Europe’s job market.
Nearly 1 in 5 European occupations now ask for AI-related skills, more than 3x the share seen in 2023.
The fastest-growing skill is not model building, but AI fluency, which means knowing how to prompt, manage, check, and apply AI systems inside real jobs.
Demand for AI fluency grew 5x, from work covering 1.9M employees in 2023 to 9.4M in 2025.
Demand for technical AI skills grew only 1.7x, which means Europe needs far more AI users than AI engineers.
The demand is still concentrated, with 75% coming from computer, management, and business-finance roles, but it is spreading into logistics, HR, compliance, and skilled trades.
Nearly 1 in 5 European occupations now ask for AI-related skills, more than 3x the share seen in 2023.
The fastest-growing skill is not model building, but AI fluency, which means knowing how to prompt, manage, check, and apply AI systems inside real jobs.
Demand for AI fluency grew 5x, from work covering 1.9M employees in 2023 to 9.4M in 2025.
Demand for technical AI skills grew only 1.7x, which means Europe needs far more AI users than AI engineers.
The demand is still concentrated, with 75% coming from computer, management, and business-finance roles, but it is spreading into logistics, HR, compliance, and skilled trades.
This media is not supported in your browser
VIEW IN TELEGRAM
Terence Tao: "We lived in a world with cognitive friction until very recently, where every task required us to use our brain.
So we didn't really think about it, we just thought this was the cost of doing something intellectual. But now we have AI and the other technologies that can bring these frictions down to zero."
Most research time is not spent having cinematic insights.
It is spent checking cases, chasing references, translating intuition into computation, testing a path, finding it false, and deciding whether the failure taught you anything.
AI changes the cost of that loop.
Terence Tao says that now he can try “crazier things,” and that makes so much difference. Because unconventional ideas are often not rejected by proof, but by inconvenience.
A mathematician may avoid a strange direction not because it is foolish, but because the bookkeeping, coding, or literature search needed to test it is too expensive for a hunch.
This is where cognitive friction becomes scientific friction.
Lowering it does not make taste, judgment, or proof disappear; it makes more weak signals cheap enough to inspect before they are abandoned.
AI is making hesitation less expensive, and that is often where discovery begins.
So we didn't really think about it, we just thought this was the cost of doing something intellectual. But now we have AI and the other technologies that can bring these frictions down to zero."
Most research time is not spent having cinematic insights.
It is spent checking cases, chasing references, translating intuition into computation, testing a path, finding it false, and deciding whether the failure taught you anything.
AI changes the cost of that loop.
Terence Tao says that now he can try “crazier things,” and that makes so much difference. Because unconventional ideas are often not rejected by proof, but by inconvenience.
A mathematician may avoid a strange direction not because it is foolish, but because the bookkeeping, coding, or literature search needed to test it is too expensive for a hunch.
This is where cognitive friction becomes scientific friction.
Lowering it does not make taste, judgment, or proof disappear; it makes more weak signals cheap enough to inspect before they are abandoned.
AI is making hesitation less expensive, and that is often where discovery begins.
✍1
🇨🇭 Global One SI - Prompt
Smart Money Tracking on Polymarket: How to Copy the Best Traders and Steal Their Edge (+ Full Guide) While most retail traders chase headlines and emotions, smart traders quietly accumulate large positions based on data, probability, and proven edge This…
This GitHub repository uses the full 31-member GFS ensemble with smart probability conversion and an 8% edge filter to trade weather markets on Polymarket
It turns 31 forecasts into one calibrated probability, applies Laplace smoothing, and enters only when the edge is over 8%
The result is a clean, disciplined dual-platform bot that scans both Polymarket and Kalshi every 5 minutes
> How it works:
Open-Meteo gives all 31 ensemble runs of GFS in one free API request
Each member starts with slightly different initial conditions, and over 24–48 hours these small differences grow into noticeably different forecasts
The spread between members becomes your uncertainty signal.
> Converting 31 forecasts into one probability - 3 steps:
Step 1 - For each of the 31 members check if the condition is met:
condition_met = forecast_temp > threshold
Step 2 - Calculate the percentage of members where the condition is true:
p = count(condition_met) / 31
Example: 28 out of 31 members show above 70°F → model probability = 90%.
Step 3 - Apply Laplace smoothing on extreme values:
p_smooth = (count + 1) / (31 + 2)
This prevents 0% or 100% probabilities from small sample sizes
> The 8% threshold
The bot enters only when |p_ensemble − p_market| ≥ 0.08
This number is not random - it accounts for taker fees (~1.5–1.8%), slippage in weather markets (~1–2%), and model error (~2–3%)
Below 8% the costs usually eat the edge
> Dual-platform approach
The same GFS ensemble and probability logic is applied to both Polymarket and Kalshi every 5 minutes
Sometimes the same event is mispriced on one platform and fairly priced on the other
The bot catches both opportunities.What you can actually take for your own bot
> The most valuable part is the pipeline:
Open-Meteo GFS ensemble → percentage calculation → Laplace smoothing → comparison with market price → entry if edge ≥ your cost threshold.
You don’t need the whole bot. This logic alone is already a solid foundation for a serious weather trading system
Repo: github.com/suislanchez/polymarket-kalshi-weather-bot
It turns 31 forecasts into one calibrated probability, applies Laplace smoothing, and enters only when the edge is over 8%
The result is a clean, disciplined dual-platform bot that scans both Polymarket and Kalshi every 5 minutes
> How it works:
Open-Meteo gives all 31 ensemble runs of GFS in one free API request
Each member starts with slightly different initial conditions, and over 24–48 hours these small differences grow into noticeably different forecasts
The spread between members becomes your uncertainty signal.
> Converting 31 forecasts into one probability - 3 steps:
Step 1 - For each of the 31 members check if the condition is met:
condition_met = forecast_temp > threshold
Step 2 - Calculate the percentage of members where the condition is true:
p = count(condition_met) / 31
Example: 28 out of 31 members show above 70°F → model probability = 90%.
Step 3 - Apply Laplace smoothing on extreme values:
p_smooth = (count + 1) / (31 + 2)
This prevents 0% or 100% probabilities from small sample sizes
> The 8% threshold
The bot enters only when |p_ensemble − p_market| ≥ 0.08
This number is not random - it accounts for taker fees (~1.5–1.8%), slippage in weather markets (~1–2%), and model error (~2–3%)
Below 8% the costs usually eat the edge
> Dual-platform approach
The same GFS ensemble and probability logic is applied to both Polymarket and Kalshi every 5 minutes
Sometimes the same event is mispriced on one platform and fairly priced on the other
The bot catches both opportunities.What you can actually take for your own bot
> The most valuable part is the pipeline:
Open-Meteo GFS ensemble → percentage calculation → Laplace smoothing → comparison with market price → entry if edge ≥ your cost threshold.
You don’t need the whole bot. This logic alone is already a solid foundation for a serious weather trading system
Repo: github.com/suislanchez/polymarket-kalshi-weather-bot
✍1
🇨🇭 Global One SI - Prompt
A new era of PC. Nvidia, Microsoft, and Arm are all teasing Nvidia’s new N1X laptop processors All eyes area now on Nvidia’s keynote at Computex on Sunday night.
Nvidia’s new ARM-based SOC is rumoured to feature:
- Up to 20 CPU cores
- Up to 6144 Blackwell CUDA cores
- Use a TSMC 3nm process
- Up to 128GB unified memory
This has the potential to be an M5 Max competitor for AI-developers.
- Up to 20 CPU cores
- Up to 6144 Blackwell CUDA cores
- Use a TSMC 3nm process
- Up to 128GB unified memory
This has the potential to be an M5 Max competitor for AI-developers.
✍1👨💻1
Microsoft open-sourced this AI quant investment platform 100% for free
Here's what it does (and how to get it for free):
https://qlib.readthedocs.io/en/
Here's what it does (and how to get it for free):
https://qlib.readthedocs.io/en/
👀1
🇨🇭 Global One SI - Prompt
Microsoft open-sourced this AI quant investment platform 100% for free Here's what it does (and how to get it for free): https://qlib.readthedocs.io/en/
Introducing Qlib: An AI-oriented Quantitative Investment Platform
✍1
🇨🇭 Global One SI - Prompt
$BB monthly chart is starting to look absolutely insane here. 👀 Most people still associate BlackBerry with old smartphones… But the market is slowly remembering that $BB now sits at the intersection of: - cybersecurity - embedded systems - automotive…
$BB
The most interesting thing here isn't the move.
It's what ISN'T happening.
No major selling pressure.
No distribution.
No heavy rejection.
Just persistent buying and volume staying elevated week after week.
The breakout above $6.95 changed the entire chart structure.
If buyers continue defending this area, the next major level sits near $12.
Momentum remains firmly with the bulls.
The most interesting thing here isn't the move.
It's what ISN'T happening.
No major selling pressure.
No distribution.
No heavy rejection.
Just persistent buying and volume staying elevated week after week.
The breakout above $6.95 changed the entire chart structure.
If buyers continue defending this area, the next major level sits near $12.
Momentum remains firmly with the bulls.
🇨🇭 Global One SI - Prompt
$AKAM what next target?
$AKAM
While everyone is focused on the obvious AI winners, some of the most interesting charts are developing quietly in the background.
Akamai has now:
• reclaimed its long-term trend
• broken above the 200-week moving average
• built a strong higher-low structure
• attracted steadily increasing volume
This isn't a hype story.
It's infrastructure.
The AI boom needs compute.
Compute needs networks.
Networks need delivery and security.
That's where Akamai sits.
One of the most overlooked charts in technology right now.
While everyone is focused on the obvious AI winners, some of the most interesting charts are developing quietly in the background.
Akamai has now:
• reclaimed its long-term trend
• broken above the 200-week moving average
• built a strong higher-low structure
• attracted steadily increasing volume
This isn't a hype story.
It's infrastructure.
The AI boom needs compute.
Compute needs networks.
Networks need delivery and security.
That's where Akamai sits.
One of the most overlooked charts in technology right now.
Forwarded from 🛢WTI, Brent Crude Oil, NGAS Trading 📊
Bloomberg reports that 29 of 109 large non-Iranian oil tankers trapped inside the Persian Gulf have now crossed the chokepoint.
That’s roughly 1 quarter of the stranded fleet.
These are big vessels each capable of hauling 700,000+ barrels.
Some supply is getting out.
Some fleet capacity is being freed.
But the system is still far from normal.
Every successful transit helps the market.
But every remaining tanker trapped inside the Gulf keeps the risk premium alive.
That’s roughly 1 quarter of the stranded fleet.
These are big vessels each capable of hauling 700,000+ barrels.
Some supply is getting out.
Some fleet capacity is being freed.
But the system is still far from normal.
Every successful transit helps the market.
But every remaining tanker trapped inside the Gulf keeps the risk premium alive.
Since next week at Computex in Taiwan will be Microsoft/NVIDIA/Arm + Qualcomm’s big week, don’t forget AMD + Intel…
🧠 Localmaxxers, the local AI silicon race is getting wild 👇
✅ AMD Halo
Strix Halo is here. Gorgon Halo / Ryzen AI Max PRO 400 is official → up to 192GB unified memory + 160GB VRAM (end of '26 w/AMD Agent PC)
🟡 NVIDIA N1X
Expected Windows-on-Arm debut with Microsoft. Still not officially announced. Official NEXT WEEK!!
🧪 Intel Razor Lake-AX
Rumored Halo-class APU → on-package memory, maybe LPDDR6, plus a much bigger Arc iGPU. end-'28
🐍 Intel Serpent Lake
Rumored Intel x86 + NVIDIA RTX tile SoC → official Intel/NVIDIA RTX SoC partnership exists, but “Serpent Lake” name/specs/timing are still leaky. Rumored late ’28/’29.
🔮 AMD Medusa Halo
Rumored Zen 6 + RDNA 5 + LPDDR6 follow-up. Rumored late ’27/’28.
💪Localmaxxing is becoming the new laptop war.
🧠 Localmaxxers, the local AI silicon race is getting wild 👇
✅ AMD Halo
Strix Halo is here. Gorgon Halo / Ryzen AI Max PRO 400 is official → up to 192GB unified memory + 160GB VRAM (end of '26 w/AMD Agent PC)
🟡 NVIDIA N1X
Expected Windows-on-Arm debut with Microsoft. Still not officially announced. Official NEXT WEEK!!
🧪 Intel Razor Lake-AX
Rumored Halo-class APU → on-package memory, maybe LPDDR6, plus a much bigger Arc iGPU. end-'28
🐍 Intel Serpent Lake
Rumored Intel x86 + NVIDIA RTX tile SoC → official Intel/NVIDIA RTX SoC partnership exists, but “Serpent Lake” name/specs/timing are still leaky. Rumored late ’28/’29.
🔮 AMD Medusa Halo
Rumored Zen 6 + RDNA 5 + LPDDR6 follow-up. Rumored late ’27/’28.
💪Localmaxxing is becoming the new laptop war.
With all the Microsoft-Nvidia-Windows-Arm N1X conversations next week, Intel's Panther Lake will have its work cut out for it in the battle for Local AI.
Intel on paper beats AMD. With up to 180 TOPS of total AI performance, it significantly outperforms AMD’s Ryzen AI 400 series (~120–130 TOPS).
Here are some advantages:
⚡ 35–47% faster local LLM inference in real-world tests
🎮 Much stronger Xe3 Arc GPU for running larger models locally
🔋 Better balance of performance + efficiency than previous Intel chips
🤖 Excellent for agentic AI, coding, and multimodal tasks
Panther Lake is currently one of the strongest platforms for local AI on laptops. 🔥 Waiting to see how N1X compares.
Intel on paper beats AMD. With up to 180 TOPS of total AI performance, it significantly outperforms AMD’s Ryzen AI 400 series (~120–130 TOPS).
Here are some advantages:
⚡ 35–47% faster local LLM inference in real-world tests
🎮 Much stronger Xe3 Arc GPU for running larger models locally
🔋 Better balance of performance + efficiency than previous Intel chips
🤖 Excellent for agentic AI, coding, and multimodal tasks
Panther Lake is currently one of the strongest platforms for local AI on laptops. 🔥 Waiting to see how N1X compares.
$FIVN
Everyone talks about AI chips.
Far fewer people talk about where AI actually gets deployed.
Customer service.
Contact centers.
Enterprise automation.
AI agents.
That's where Five9 sits.
Now look at the chart:
• multi-month trend reversal
• higher lows developing
• volume expanding
• momentum improving
• reclaiming key moving averages
The strongest trends often begin long before they become popular.
This is one of the more interesting software charts on my screen right now.
Everyone talks about AI chips.
Far fewer people talk about where AI actually gets deployed.
Customer service.
Contact centers.
Enterprise automation.
AI agents.
That's where Five9 sits.
Now look at the chart:
• multi-month trend reversal
• higher lows developing
• volume expanding
• momentum improving
• reclaiming key moving averages
The strongest trends often begin long before they become popular.
This is one of the more interesting software charts on my screen right now.
$MLGO
One chart that keeps showing up on my radar.
After months of compression, the stock is now pressing directly against a major descending trendline.
What stands out:
• volatility contraction
• improving structure
• volume starting to return
• reclaiming key moving averages
• breakout attempt developing
The move hasn't happened yet.
But that's usually when the best risk/reward opportunities appear.
Worth watching closely next week.
One chart that keeps showing up on my radar.
After months of compression, the stock is now pressing directly against a major descending trendline.
What stands out:
• volatility contraction
• improving structure
• volume starting to return
• reclaiming key moving averages
• breakout attempt developing
The move hasn't happened yet.
But that's usually when the best risk/reward opportunities appear.
Worth watching closely next week.
$FJET One chart quietly getting stronger while nobody is paying attention.
After months of consolidation, FJET is now pressing directly against a major descending trendline.
What stands out:
• higher lows forming
• volume returning
• volatility compression
• improving momentum
• aerospace and space themes are hot
The move hasn't happened yet.
But the best setups usually look boring right before they become obvious.
Worth watching very closely next week.
After months of consolidation, FJET is now pressing directly against a major descending trendline.
What stands out:
• higher lows forming
• volume returning
• volatility compression
• improving momentum
• aerospace and space themes are hot
The move hasn't happened yet.
But the best setups usually look boring right before they become obvious.
Worth watching very closely next week.
$KITT One of the most compressed charts on my screen right now.
After weeks of consolidation, KITT is now pressing directly against a tightening triangle structure.
What stands out:
• sellers disappearing
• volatility compression
• higher lows forming
• volume drying up
• defense and robotics themes gaining attention
The biggest moves often start when nobody is paying attention.
This chart looks very close to a decision point.
Worth watching closely next week.
After weeks of consolidation, KITT is now pressing directly against a tightening triangle structure.
What stands out:
• sellers disappearing
• volatility compression
• higher lows forming
• volume drying up
• defense and robotics themes gaining attention
The biggest moves often start when nobody is paying attention.
This chart looks very close to a decision point.
Worth watching closely next week.
SoftBank just pledged €75B to build Europe’s largest AI computing facility in France, turning cheap, stable nuclear-heavy power into the raw fuel for training and running bigger AI systems.
The plan starts with €45B for 3.1GW of capacity in Hauts-de-France by 2031, then adds another 2GW, creating a 5GW AI complex roughly comparable to the power draw of a major city.
Dunkirk is the anchor site, with Schneider Electric expected to help build an AI infrastructure and robotics manufacturing hub around the project.
The weak point is financing, because industry estimates put 1GW of AI infrastructure near $50B, meaning SoftBank will probably need heavy project debt and outside partners.
The plan starts with €45B for 3.1GW of capacity in Hauts-de-France by 2031, then adds another 2GW, creating a 5GW AI complex roughly comparable to the power draw of a major city.
Dunkirk is the anchor site, with Schneider Electric expected to help build an AI infrastructure and robotics manufacturing hub around the project.
The weak point is financing, because industry estimates put 1GW of AI infrastructure near $50B, meaning SoftBank will probably need heavy project debt and outside partners.
ByteDance is building its own AI data-center CPUs because running agents at TikTok scale now depends on scarce server processors, not only Nvidia GPUs.
inspired by Groq's "language processing units," they are testing both Arm and RISC-V, which lets it compare a mature commercial design against a more controllable open instruction set before mass production.
The market is seeing a 10%-35% quarterly CPU price increases and long supply delays, hence making an in-house silicon is now cost and supply-chain move, not just a prestige project.
So ByteDance wants to both reduce dependence on restricted foreign AI hardware and make inference cheaper per query.
The deeper shift is that AI agents is now turning CPUs into strategic chips. A
gentic inference stresses CPUs much more because one user request can trigger many smaller steps: retrieve files, call a tool, query a database, run a model, check the answer, call another model, send data across servers, and manage memory.
However, ByteDance does not seem to have in-house chip design teams and is reportedly relying on several external partners, who are also expected to handle the actual silicon manufacturing.
inspired by Groq's "language processing units," they are testing both Arm and RISC-V, which lets it compare a mature commercial design against a more controllable open instruction set before mass production.
The market is seeing a 10%-35% quarterly CPU price increases and long supply delays, hence making an in-house silicon is now cost and supply-chain move, not just a prestige project.
So ByteDance wants to both reduce dependence on restricted foreign AI hardware and make inference cheaper per query.
The deeper shift is that AI agents is now turning CPUs into strategic chips. A
gentic inference stresses CPUs much more because one user request can trigger many smaller steps: retrieve files, call a tool, query a database, run a model, check the answer, call another model, send data across servers, and manage memory.
However, ByteDance does not seem to have in-house chip design teams and is reportedly relying on several external partners, who are also expected to handle the actual silicon manufacturing.
🇮🇷 Every single stock on the Tehran Stock Exchange closed green today. Not one in the red.
First time in the exchange's history.
Iran's market is pricing in the deal before the ink exists. When a population that's been cut off from the internet for 5 months sees every stock go green on the same day, they know something is coming.
The Supreme Leader hasn't signed. The terms are still disputed. The market already decided.
Source: Middle East Spectator
First time in the exchange's history.
Iran's market is pricing in the deal before the ink exists. When a population that's been cut off from the internet for 5 months sees every stock go green on the same day, they know something is coming.
The Supreme Leader hasn't signed. The terms are still disputed. The market already decided.
Source: Middle East Spectator