rjurney

Sounds like he ran up against the snails pace of enterprise sales. It takes patience. When I cofounded a company selling a KYC solution to global banks, I did a survey of 30 FinTech founders on how long it took to get ink on paper with a global bank. 18 months was the usual answer, and it took even longer to get an actual check. If demand for your product is from large enterprises and you don't plan for this up front you simply can't survive. SaaS and "no meetings" are a great alternative... if the demand is there and it scales to a real opportunity. A lot of startups get lured into dealing with calls because a huge company with a potential $1M+ sale looms and they could raise their next round now if they close it. It is hard to say no.

No, but I've pushed myself to the brink for two years, burnt out from a failed startup and then taken two years of sick or half-sick time to be healthy, productive me again. Take the time you need before this happens to you. If you can afford it... do it!

It's kind of interesting in terms of analytics... can we predict when lenses will appear and disappear, from our perspective? What might we do with that information once we are more advanced?

Downloaded for future RAG / LLM retrieval :)

It's just an unfortunate title by her editor, the article is interesting. Silicon Valley is about building new companies, not large incumbents manipulating the world.

My question is... we still spend a ton on defense. How did the budget change?

What an awesomely comprehensive resource, reminds me of MITRE ATT&CK.

Thanks, I figured but do you know just how much larger? I am listening to Ken Burns' Vietnam on Audible and I believe we had 200K personnel in Vietnam at the peak... oh shit I just double checked and it was 543K personnel JUST IN VIETNAM. Thanks, mang! Enlightening. That is 25% of our entire military today.

I did not miss the point of the article, I just don't get his interest in medium helicopters when the missile he wants 8 of on a medium helicopter weighs 100 pounds.

We took out tons of infrastructure in Iraq during shock and awe. Utilities were on the target list. We were about to occupy it. That was incredibly stupid. The infrastructure itself was not collateral damage, it was targeted. We have no occupation plan, it was that stupid. The destruction resulted in millions of extra deaths due to the impoverishment and destruction of Iraqi society. Yes, we bear responsibility for all those deaths. You break it, you own it. That's war.

Oh they will. They are the sensor-fused quarterbacks of drone warfare.

When you topple a foreign government, destroy all the infrastructure for pointless "shock and awe" and then send the ethnic majority but recently oppressed armed forces home... you bear responsibility for the millions of extra deaths that follow when traumatic civil war rocks the nation. You are the exact example of the delusional American he means.

Flying down canyons was so much fun, like a trench run.

Well, actually helicopters are already critical in both anti-submarine and anti-ship warfare. They can extend range and distribute shots of anti-ship missiles to make concurrent salvos from multiple directions more deadly.

They stall and aren't maneuverable at the speeds you need.

Brimestone 2 is 50kg - what's his deal with medium helicopters?

This seems to be the takehome. The helicopter has to be a drone operator and behave very defensively, with effective active measures.

> Given the right network integration and the right weapons, you could inflict a deal of pain in a fairly basic helicopter while staying well out of the way of any air defence. You still have to protect yourself against chance encounters – partisans or special forces with shoulder-launched weapons, for example. This appears easier said than done.

> There were about 11,846 U.S helicopters that served in the Vietnam War. The U.S records show 5,607 helicopter losses.

Are 310 helicopter losses in Ukraine that high? I don't think so. Russia has lost =~ 3K tanks. That's a 10:1 ratio for tank:helicopter, is that high?

Helicopters' role may change. They would be ideal drone operator platforms - they already have a weapons operator in a back seat. The right cannon / loadout could take out small drones. Collaborative Combat Aircraft (CCA) style wingmen would work with quadcopters for helicopters. They could expand the sensory capability to avoid surface to air threats.

But there must be many more of the one than the other, since one has to satisfy a constraint and the other does not.

I got this one without realizing the complex basis for it... there's an infinite number of real numbers between 0 and 1, compared with a limited proportion of rational numbers.

Thank you for taking the time for this thoughtful reply. I did not mean literally posting works. What I meant was I do get followers for my link posts but... actually one where I posted just a photo did 10x better than any of my previous posts. The best posts I've done have been LinkedIn Blog Posts.

I'm not an expert at this, and you are, so tell me... should I really not use images? Like, I often post a link to a paper and then I will summarize it and rather than a link I will post one of the figures from the paper. What's a better strategy than that?

Computer science. More and more researchers are on LinkedIn, and if you post a paper and @ their account, you are instantly on their radar since most are never recognized like that. Pro tip ;)

You fucks, I just learned about this. I feel so stupid!

Twitter is how I built my entire network in the Bay Area... it was amazing at flattening hierarchies, I could talk to anyone. Then at some point the recommender system changed and people in my network never saw tweets from one another. At that point it became useless for my career.

Concurrently, LinkedIn sharing grew from nothing to awesome. I get most of my news about the latest developments in my field from LinkedIn. I get jobs by posting blog posts on LinkedIn, or summaries of interesting recent papers.

Yes, and I don't get it. Why would people read a post without a link? My link posts do fine.

I have no trouble sharing links. Even multiple links in one post, although the first one will define the preview. Nor do I have a problem sharing images and including urls in the post body. Nor do I understand why people only share the link in the top comment.

What in the hell is he talking about? It is a good idea to include some quotation or a summary of the link's content. Other than that, how is sharing links painful?

I think LinkedIn is fantastic. Early in my career my LinkedIn resume was how I got work. At this point, sharing links and my own blog posts to keep my 6,500 followers interested in me as an expert is how I get work.

The only part I don't like is sharing a Post to All in a group... one group at a time. Sharing in a few groups is a good way to drive views but I've developed excellent motor skills at navigating all those clicks necessary to do it.

The reason they have so much power on the supply side is that what they do is incredibly painstaking and expensive. I agree the system sucks.

Sublingual Toradol is available from pharmacies in Mexico. I fly to San Diego every five years, take the tram to the border and walk across to Tijuana to stock up. Toradol is an NSAID they give you via injection in the emergency room for chronic pain and headaches. By mouth it tends to cause GI bleeding worse than any other NSAID. A number of people died before they restricted the pill to five days consecutive use. Dissolving Toradol under your tongue bypasses the GI tract, making sublingual Toradol much safer. I carry a pill pack of four pills in my wallet, and having a pain medication as effective as morphine without sedation in my pocket at all times seriously improves my quality of life. There is something I can DO when my chronic pain flares up enough that it starts to get to me.

In the US the Sprix nasal inhaler is the functional equivalent of sublingual Toradol. I've read the research papers on both medications and I compared the charts for pharmacokinetics and pharmacodynamics and... the curves are virtually identical. They have the same impact on the body, both delivery mechanisms are equally effective in safety terms.

Sublingual Toradol is generic - the original papers on it are from the 80s. The Sprix is a patented delivery mechanism... wait for it...

* A dose of subligual Toradol in a Tijuana pharmacy: $0.25

* A single dose Sprix nasal inhaler from a US pharmacy: $25 last I checked. Only indicated for severe migraines.

That's a price multiple of 100x, as in one hundred times for a functionally equivalent medication. Last time I went to Tijuana the murder rate had multiplied by 10x since the last time I was there, so I went to the pharmacy right at the border. The price per dose was $0.50 there, but its cheaper if you go downtown. It isn't a counterfeited medication, so you don't have to worry about that. I wish more pain patients had access to sublingual Toradol. If a drug company wanted to improve chronic pain care, they could fund clinical trials in the US and get the medication approved. It will never happen in the states. Pure greed is the ONLY reason.

Actually, there's a lot more research on sublingual Toradol since I last checked: https://pubmed.ncbi.nlm.nih.gov/?term=%22sublingual+toradol%... See the sublingual film work? No reason for that instead of a pill except you can patent it. BOOM. BURN IT ALL DOWN.

I think it depends on what your use case is and how optimized your prompts are.

Which benchmarks are you looking at? It is very competitive with GPT4o in the table of metrics I just built at work. Have you used it to code? Qualitatively, it is much better - once it can execute Python it will be supzors.

"Open up CUDA" - guys, its all open source. What do you want them to do? Do tech support to help their competitors compete against them? AMD is to blame for not building this project 10 years ago.