condiment

To respond to the 'nuance' of your remark, that helmets change rider behavior for the worse, resulting in higher aggregate injuries - that is also incorrect. The passage of helmet laws results in significant reductions (20-50%) in head injuries and deaths. These are reductions among the same population, in the same geography, in a short timeframe. It is indisputable.

If the total number of recorded injuries is going up, it's because ridership has increased. Ridership is up for lots of reasons, population growth and health benefits being two of them. Cycling is a terrific way to improve your overall health, even when the risk of injury or death due to cycling is taken into consideration. [2]

And if manufacturers profit from improving the health and safety of a population? Good.

[1] https://newrossgreenway.org/bicycle-helmet-vs-no-helmet-stat... [2] https://pmc.ncbi.nlm.nih.gov/articles/PMC10546027/

Consultants are hired to improve decision quality. They do that through a combination of analysis, experience, and industry connections. The work product is rarely 'powerpoints'. The work product is the actual work, which entails people doing research, making decisions, executing on them, and measuring results.

At any rate, I don't disagree with you that consultants aren't going anywhere. This post is an ad for an immature AI service that makes a bunch of claims about how they've solved for hallucinations and knowledge cutoffs in order to take down McKinsey. If they had solves for those problems, they'd be wasting their effort building a replacement for consulting firms.

That's not to say there aren't additional ethical challenges that would arise if gene therapies were cheap, but the ethics concerns you're raising seem like future concerns, relevant to a world that does not yet exist.

That approach developed the technologies, but what if a mass market approach is what is required to create the economies of scale we need to actually benefit from this science long-term? Consider a cure for the common cold. All of the technology exists. We can take a tissue or fluid sample and extract viral genomes, comparing against a database of over 100k viruses. We can develop MRNA vaccines that target specific viral proteins. We can do both of these things using current technology in a timeframe of less than a day.

To me this is the next great leap. Gene therapy is less a technology problem than it is an infrastructure problem, and rare diseases cannot support the infrastructure. But there are plenty of wealthy people who would pay $10k, maybe even $50k, out of pocket to cancel a cold. Rare diseases bootstrapped the tech, common diseases can scale it, and once scaled, everyone benefits.

You are correct there’s no problem with them if they’re used enough, but evidence suggests they do not receive that usage.

If a $10,000 investment reduces the chances of a serious accident by 90%, the corresponding reduction in insurance rates might have a payoff within a few years. Especially if adoption starts to push rates up for customers who don’t automate. I can’t take a taxi everywhere, but I’d sure like it if my car drove me everywhere and did a better job than me at it too.

There are some experimental specifications that exist to provide attestation as to the authenticity of media. But most of what I’ve seen so far is a “perjury based” approach that just requires a human to say that something is authentic.

    total_duration:       10530451000
    load_duration:        54350253
    prompt_eval_count:    36
    prompt_eval_duration: 29000000
    prompt_token/s:       1241.38
    eval_count:           460
    eval_duration:        10445000000
    response_token/s:     44.04

Anecdotally I can share my revealed preference. I have both an M3 (36gb) as well as a GPU machine, and I went through the trouble of putting my GPU box online because it was so much faster than the mac. And doubling up the GPUs allows me to run models like the deepseek-tuned llama 3.3, with which I have completely replaced my use of chatgpt 4o.

https://drs.faa.gov/browse/excelExternalWindow/DRSDOCID17389...

I’m not sure this makes sense in 2024. Pretty much every web server supports websockets at this point, and so do all of the browsers. You can easily impose the constraint on your code that communication through a websocket is mono-directional. And the capability to broadcast a message to all subscribers is going to be deceptively complex, no matter how you broadcast it.

This is a really cool example of ambient technology. Typically when people talk about ambient technology they're talking about something like an e-ink display that is pushing information to you, but in a way that doesn't require interactivity and isn't screaming for your attention. This is a little different in that it's always _receiving_ information from you without any need for interaction or maintenance, except on your terms.

The interaction model is pretty clever too. Since it's collecting data from the instrument they've found a way to cue the device to perform an action without the user needing to open an app. (black keys to bookmark) There is an app of course, but it connects directly to the device, with no annoying setup requirements. I've seen this same approach with several other devices - Xbloom coffee maker, Combustion thermometers, Week Aqua lights - it works really well. I'm understating it. It's astounding how pleasant it is to use devices like this.

As hardware continues to improve I expect this will be the default mode for pretty much every new technology appliance. Ambient operation, local data, local app, with cloud and accounts as _options_ to extend functionality if it's necessary.

In an organization where people are accustomed to indexing their activities to other people who live in a different time zone, it’s actually easier to use the most common time zones than it is to switch everyone to UTC. When using UTC, you have to do the mental gymnastics twice. You first have to relate your own time zone to UTC, and second relate your audience time zone to it. But what happens in practice is that you quickly learn what times in your local time zone correspond to movements of the day in your peer time zones. And because everybody has the same mental model, it’s easier (and less error prone) to just to use local time zones.

The truth is that only some cable companies make these investments - you can look up “fiber node size” for respective performance across different companies. A fiber node being the place where optical is terminated and switched to coax. These have been getting smaller everywhere, but it only makes sense to invest there when the upstream infrastructure can support it. So from a consumer perspective, your “Linux isos” will be slow to download in any case until the upstream network is upgraded and your node is split to offer higher performance.

These rates are independent. You can’t aggregate them because the level of effort to secure VC funding is very different from the level of effort required to take a funded start up to unicorn heights.

If I understand your argument to be that startups are relegating themselves to the trash bin because they’re attempting to get attention from venture capitalists at the expense of some fundamental “goodness”, some data supporting that point would be helpful. But that’s not what’s in the article.

^^ Here's an article written by Lotte Knudsen, referenced in the original post, that further tells the story of how GLP-1 was first developed into a drug (as liraglutide) and approved for human use. There were a lot of false starts and additional practical problems that needed to be solved in order to yield a viable medication.

After reading the OP I was surprised to learn that Novo Nordisk picked up the research only a couple of years after GLP-1 was abandoned by Pfizer, after which it took 5 years to develop the initial medication and another 12 years to make it through FDA approval. Even after all that, the primary indication was for diabetes. It took another 7 years for semaglutide to make it through approvals and bring GLP-1 into the public consciousness.

When you consider the amount of time involved, and the sustained investment required, it's difficult to fault the execs at Pfizer for their decision to shut the project down. Obesity wasn't nearly as prevalent then as it is now, and it seems likely they had funded the startup specifically because of the author's prior research into nasally-administered meds. It's even possible that the shutdown decision had little to do with the primary area of research.

As it is the only meaningful conclusion to draw from this data is that we need more data.

For me this is one of the strongest motivators for running LLMs locally- even if they’re measurably worse, they’re a far better tool because they don’t change behavior over time.

You make http requests to the shared server, those get proxied via the ssh tunnel to his machine, and the client on his machine could make the determination when/whether to run the workload.

For reference, 500 miles is the distance between San Diego and San Francisco. That’s a 90 minute flight that would take closer to 2 1/2 hours with this air taxi. So I suspect that this demonstration is less about any sort of aspiration to replace batteries and fossil fuels with hydrogen, and more a demonstration of an operationally clean burning fuel.

The criticisms that hydrogen is environmentally expensive are valid, but what’s interesting is that the operational use of the vehicle, whether it’s battery or fuel cell, is effectively emissions-free. When the operational fuel is clean, emissions improvements can be centralized. That’s a really big deal.

https://www.jobyaviation.com/news/joby-progresses-next-phase...

The interesting mechanism there is that you can buy a lot of seats at once, but you don’t get to choose where they are exactly, only the section. So in every case you’re going to have people buying big lots of tickets and distributing them to friends and family after the fact.

Then: Support a single platform, Windows or Mac

Now: Support Windows, Mac, and Linux, ARM or x86, as well as mobile.

Then: Support a single market

Now: Support a global market with region-specific localization

Then: Write software for wealthy and highly technical early adopters

Now: Write software that anyone in the world could possibly consume

Then: No accessibility

Now: Accessibility requirements for software being codified into law

Then: Zero or low consumer expectations, win by creating something new (blue ocean, lotus123 versus literal paper)

Now: Nearly everything has been done, win by differentiating on comparative quality in a field of options (red ocean, figma vs keynote)

https://www.noaa.gov/news-release/noaa-forecasts-above-avera...

Edit found the historical data: https://cdn.coastalscience.noaa.gov/page-attachments/Forecas...

Which is why it's strange to see people surprised that their $3,800 CB2 sectional isn't built to the same standard of quality as a $13,000 Roche Bobois sectional. Especially when they could have easily gone to the showroom and lifted a section up by a corner to learn exactly what they were buying. Or gone online and searched for clones, or called up any interior decorator in the region and just asked "where do I get a good couch?"

Manufacturers don’t get a bye from safety just because they haven’t sold many units yet.

Teams? I'm not using a persona in a professional setting. Also MSFT's implementation has really low information density. Slack? I need to be able to copy and paste, and the ipad compatibility app stinks. Outlook? Again, the ipad compatibility app is no good. Excel? Give me a break. I had tried the browser first, but it turns out that I have keyboard workflows that are very efficient, and requiring my gaze to shift is no good. To add to it, the gaze tracking highlights controls all over the app when I'm just trying to type, and if I glance at another window for reference, shifts my keyboard and mouse attention.

Multiple monitors projected from a real computer really would be ideal here.

On a five-year time frame, section 174 is equivalent to the existing tax code. You have five years of labor expense, and after five years 100% of it is amortized from a cumulative point of view. So the look of it is that the tax code is favoring companies that are established and survive, which doesn’t seem like so bad of a thing. If your start up is operating at a loss and you don’t have the capital to afford taxes while you’re operating at a loss, It’s a harder road to climb. So from an economic stability point of view, it makes sense to me that you would want startup companies to be well funded and for them to have a business model that doesn’t rely on a tax loophole to cannibalize existing aspects of your economy.

The foreign labor deduction Requiring 15 years of amortization versus five years for local labor is an interesting one too. That is a very clear shot at offshoring and I’d guess is the reason this wasn’t proactively revised.

Most of the analysis is “look at my million dollar ARR business and how bad it is now” but without a time component, that’s a disingenuous take.