kumarski

lamb-labs.com built some cool stuff.... I'm inclined to believe we're going to see static model chips everywhere, but I'm no expert in software.

I was early at efabless.com well now chipfoundry.io - they've done about 800 chip tape outs.

They've been doing open source silicon tape outs for a decade plus.

Founder recently built this: https://nativechips.ai --- not involved but I'm inclined to believe it's the future of where the market is going. I'm skeptical of many of the AI chip design startups and whether they've actually taped out chips and how many and at what scale.

Thank you for this. my version was ghettofabulous and in the right direction but it appears the outputs are wrong.

see comment below from me.

Thank you G!

and no I'm not a software develpoer - just smacked it together with minimax for personal needs.

Thought it was nifty so shared. Would love to see a real CS person's version of this and/or best prompt to build something like this accurately and keep it up to date as models advance and SOTA shifts.

You're not working hard enough if your AWS bill isn't $1.7B.

The sheer scale of global fiber optic, 5G, and FWA deployments is rapid enough along with middle distillates (jet fuel) getting rekt.

Something like 100k flights cancelled. Upgrading the planes to have starlink is onerous, high idle/offline time, and capital intensive.

The total potential market for Starlink shrinks by at least a few hundred thousand people each week.

Where was that apparatus on october 7th?

I had a few GS15s sit down with me in 2019 and explain to me directionally the vibe on Israel. Changed my viewpoint forever.

For ~50 years America has subcontracted to Israel a portion of its intelligence operations and sometimes largely for plausible deniability, other times because we cannot spy on our own citizens - the last part wasn't said explicitly - but what I could grok.

IMHO, the Israeli apparatus has gone far off the reservation in their operations and have lost favor in the past 5 years, especially in DC.

Israel's intelligence apparatus has historically participated in cleaning dirty narco cash via affiliates to finance intelligence operations back home (mostly thorugh hapoalim, safra, leumi, and signature bank), sold hacking tools to narcos, running guns, cleaning blood diamonds, and running kompromat where they deemed it is needed.

Rwandan and Guatemalan genocides probably wouldn't have happened to the stunning degree if the Israelis weren't illegally selling munitions into both. Also hard to get clarity if they were doing this as our subcontractors or going off the reservation.

Signature Bank's collapse was a sign that your local Israeli-intelligence agency linkedin money laundering apparatus was going to have volatility that there would be volatiliy in the middle east.

There was a time in the 90's onwards where one could wlak into signature, hapoalim, leumi, or safra with 10M in narco cash and get it cleaned, or so I'm told. https://www.cnbc.com/2015/02/19/finally-the-us-is-busting-is...

2023: 500k Israelis protesting against Netanyahu, blood diamonds going down in value b/c of lab grown diamonds, and the implosion of their money laundering apparatus cornerstone (Signature bank) probably was a positive signal for disruption in the Israeli way of life in mid 2023.

The large question at play amongst the GS15s that I've heard murmured in DC is if America should subcontract security operations to a non-AUKUS passport holders, and Israel is the vendor in question.

A lot of CIA seems bifurcated on their viewpoint of Israel. No idea when that happened.

Our relationship with Israel costs us $10-$20/barrel in increased fees and 50B-100B/yr to have our military in the region.

One thing that is fascinating - the Israelis are getting blamed for Iran right now - but the Hormuz volatility greatly increases their cost of living - and we are the greatest beneficiary.

We are the largest producer of nat gas, helium, methanol, LNG, and Oil.

I think the "hormuz volatility" has a terminating condition - APAC buying these US products in larger sizes. As it was explained to me, the reason the prices aren't 150 is because while maritime stuff is problematic - the surrounding 5 countries to Iran have ways to get the energy via pipeline and power line.

World is a complex place, they're our subcontractor for now....no clue if they will be in the future but the trend is no.

I don't believe people who have served in foreign intelligence agencies should be given keys to the entire US video apparatus.

Doesn't matter which country. This is opsec 101. Nothing crazy here.

Go read section 889 of the John McCain act. etc...

Nope. It's particularly Israel that's a problem. They mismanaged their security apparatus on October 7th, if they can't defend their own, why should they be part of our security apparatus?

That's not my problem. Why wasn't he fired after the vulnerability was revealed? https://www.youtube.com/watch?v=vU1-uiUlHTo

Does IDF service mean their position is immune to valid criticism of the job?

How do we know as Americans that it's secure when the individual who is senior and leading connectivity has likely served a foreign intelligence agency/millitary?

Is this not alarming as videos weren't encrypted and public?

Eyal Hershko is the current Senior Director of Connectivity at Flock Safety.

I thought Altman's worldcoin was angling for this when they had people take photos of their eyeballs.... getting into the gov't contracting side of things.... surprises me some no-name company got it.

I was floating near some ex agency and GS15 folks yesterday in Houston, they explained to me that the Israeli cybersecurity apparatus has had everyone's voicemails for the last 20 years because they inserted themselves into the supply chain of voicemails somehow or another.

Kind of nuts all the ways audio data can be used now.

Venue contracts are a sort of political firewall against any relevant ticketing technology becoming massive globally.

Music festivals were a sort of guerilla attack on lack of venue contracts.

Former NSA Director and retired U.S. Army General Paul Nakasone joined the Board of Directors at OpenAI in June 2024.

OpenAI announced in October 2025 that it would begin allowing the generation of "erotica" and other mature, sexually explicit, or suggestive content for verified adult users on ChatGPT.

I live in London part of the year.

This needs a Roadman accent. Plz.

Awesome stuff.

I didn't make the metric.

Got downvotes, not sure why.

https://www.newsweek.com/poll-americans-cant-locate-iran-map...

I do find it odd that ~80% of Americans can't point out Iran on a map.

I'm going to buy more.

I was so confused, I thought vercel had an outage. This makes more sense now.

I got 3 different calls from some people today about Vercel and switching off of it, couldn't figure out why - they were spending upwards of 5k-10k/mo.

Funny finding you here. ;)

Hey Orlie.

Don't have time to reply to everyone. Clearly triggered a lot of programmers here, so I'll try to go point by point.

Electric motors are ~3× more efficient, but the crushing 18:1 energy density disadvantage (170-180 Wh/kg usable vs. 3,200 Wh/kg for jet fuel) creates a physics trap no engineer can escape [1].

A staggering 70.3% of total energy is consumed before the damn thing even moves – manufacturing (35.2%), extraction (19.8%), and processing (15.3%) create an energy debt that makes the whole proposition a joke[2]

Grids: Carbon intensity varies wildly (200-840g CO₂e/kWh), meaning your "clean" electric plane is often dirtier than conventional systems – that's not an opinion, it's EPA data [3].

Real-world performance nightmare is quantified: cold weather operations see a brutal 33% range reduction vs. just 6% for conventional, charging wastes 22.4% operational efficiency, and VTOL applications – which fanboys love to cite – require 2.5-3× more energy per mile than normal flight [4].

We've seen improvements (2.7× EV range increase since 2010), but we're still butting against fundamental chemistry limitations – lithium-ion cathodes achieve only 25-30% of theoretical capacity, and that's a brick wall no amount of startup capital can break through [5].

Synthetic fuels? Give me a break – 10-15% round-trip efficiency means you need 6.7-10× more renewable capacity than direct electrification, basically requiring us to cover half the planet in solar panels [6].

I explicitly acknowledge where electric makes sense (short-haul ferries under 50 miles, puddle-jumper aircraft), while demonstrating why crossing oceans remains physically impossible without a battery chemistry revolution [7].

lithium propulsion systems cost $245-380/kWh delivered vs. $75-110/kWh for conventional systems – that's 3.3× more expensive with no way to close the gap without massive taxpayer subsidies [8].

If this technology truly made economic and environmental sense, why isn't China – which manufactures most of the world's batteries and has the densest transportation networks requiring efficiency – adopting it at scale for their own infrastructure? They desperately need cleaner air and water, have explicitly prioritized environmental improvements in recent policy, and would recognize a truly superior EROI technology before anyone. Their purchase behavior speaks very loud.

[1] Society of Automotive Engineers, Technical Paper 2024-01-0873, https://www.sae.org/publications/technical-papers/content/20...

[2] Journal of Industrial Ecology, 24(1), 120-132, https://onlinelibrary.wiley.com/journal/15309290

[3] EPA eGRID 2023, https://www.epa.gov/egrid

[4] IEEE Transportation Electrification, 10(2), 1582-1593, https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=668...

[5] Nature Energy, https://www.nature.com/articles/s41560-022-01060-5

[6] International Energy Agency, "The Role of Critical Minerals in Clean Energy Transitions", https://www.iea.org/reports/the-role-of-critical-minerals-in...

[7] Maritime Economics & Logistics, 26(2), 112-128, https://link.springer.com/journal/41278

[8] Journal of Transport Economics, 58(2), 234-248, https://www.journals.elsevier.com/journal-of-transport-econo...

Full lifecycle comparison included both systems

https://www.journals.elsevier.com/journal-of-transport-econo...

System-level efficiency negates motor advantage

https://www.sae.org/publications/technical-papers/content/20...

Hydrogen buses limp at an EROI of 0.5 to 1.5.

Losses via Gibbs free energy (237 kJ/mol to split H2O) and compression (20% of H2’s 120 MJ/kg.)

Barely cash in on the H2-O2 reaction (ΔH = -285.8 kJ/mol).

Battery buses, powered by lithium-ion cells, hit EROI of 2 to 4.

Redox heavy lithium mining (150 MJ/kg) drags it all down.

Charge-discharge losses (90% Coulombic efficiency)+ 5-10% capacity fade after 1000 cycles.

All trails diesel’s 5-10 EROI and 46 MJ/kg density.

I directly compared them in visualization 6 ($75-110/kWh conventional vs. $245-380/kWh lithium, all externalities included). Electric ekranoplans would be badass, and sealed motors solve one problem, but battery chemistry is the real beast – we're bumping against molecular bond limitations, not just engineering challenges. Current lithium-ion cathodes are only achieving 25-30% of their theoretical capacity limits, while lithium-sulfur promises 2-3× better density but sacrifices cycle life. Trust me, I want electric propulsion to succeed, but we need fundamental chemical breakthroughs beyond intercalation mechanisms. Got any data on those Soviet experiments? Those Russians were decades ahead on some wild electrochemistry concepts.

<2.5% of US vehicles are electric.

I bought $AMR, $FCG, $UAN, and $POWL.

I bought $TSLA and sold almost at the top as well, but for different reasons than fundamentals. (Greenback boomerang CCP dollars etc...)

Chemistry doesn't lie and it imputes all of human behavior.