faizanbhat

Hi all, I wanted to share a plugin that I wrote for Claude code. I’ve been working full time with Claude code over the last few months. I’m generally a big fan and have been using it for product thinking and development. I found it frustrating that important product context that I built during sessions vanished after compacting or switching sessions. I built manual workflows to work around this, essentially structuring and storing product context like strategy docs and mental models. This plugin is an attempt to automate and streamline what’s worked well for me. I’d love to hear feedback and suggestions. Thanks!

The U.S. has many tactics at its disposal:

1. Sanctioning bad actors: Placing wide ranging economic sanctions on bad actors is a potent tool but it can backfire. U.S can penalize any company that does business with sanctioned individuals. In the case of China, applying sanctions on party members would make it virtually impossible for them to transfer their wealth overseas via global banks, property markets, investment vehicles, etc. This ratchets up the pressure on the Chinese government as it immediately and adversely affects the interests of China’s powerful elite. The downside of this approach is that China is likely to retaliate against U.S. economic interests within China. It’s a large market coveted by many U.S. companies, so there is likely to be political blowback, which makes this unlikely to happen.

2. Diplomatic pressure to isolate China: China cares deeply about how it’s perceived on the world stage. We rarely hear strong international condemnation of China’s social, political or economic policies. This is partly due to the China’s success in using their economic power to strengthen their global standing. Much has been written about China’s debt diplomacy, for example. China now plays an outsized role in organizations like the WHO and various UN bodies. It’s even a member of the UN human rights council. The U.S. on the other hand has been withdrawing from these bodies, effectively ceding the stage to China. The U.S could apply pressure on China by once again assuming its leadership position within these bodies, and working with allies to counter Chinese influence and condemn China’s internal and external policies. China has no effective response to this tactic and it’s therefore one that they are particularly concerned about IMO.

3. Stronger military and economic alliances with Taiwan, India, Japan and Australia would create a counter balance to rising Chinese dominance in the region.

4. The U.S can also take steps to prevent knowledge transfer to China by limiting foreign student intake, or preventing research collaboration with Chinese universities.

Looks awesome. Bookmarked!

The fact that I worked in ad tech doesn't automatically make an advocate for ad targeting. I don't like being tracked either.

As to your question on why it's being done if it isn't effective, I believe there are two reasons for it:

1. A vast % of digital ad spend goes to a few companies that have personally identifiable information and very effective tracking. Google, FB. I did point of this exception in my comment. Frankly, these companies don't need to target you to be effective. They gather their richest data from your usage on their service.

2. As for the rest of ad tech: Media buyers at agencies tend to be fresh out of college grads. They don't fully grasp the capabilities/limitations of ad tech products, targeting effectiveness, etc. Agencies want to spend their digital budgets and not exclusively on FB, Google. They want to show that they're diversifying, innovating, etc. This IMO is how the rest of ad tech stays afloat.

Amazon, Ebay, Walmart, etc. have persistent cookies. They are mapped to your profile in their database. If you delete the cookie, it reappears the next time you sign in. These companies can collect vast amounts of data on you, which allows their targeting to become highly effective over time.

But these companies are few and far between. Most ad tech companies have nowhere near this quality of data or access to users.

As someone who worked in ad tech for half a decade, I fail to understand the fuss about cookie based targeting. We're speaking of databases mapping data to random numbers stored in a file that can easily be deleted. The real issue is ad targeting by a handful of large companies that have personally identifiable information.

It's not like cookie based targeting is very effective either. It's very difficult to get cookie based ad targeting right – you have to make significant investments into product, technical integrations and data buys. Given the fact that the cookies to which you are tying these investments can vanish at any time (and a large % do so regularly), it rarely makes business sense to make the right sort of investments. Instead most companies do the bare minimum and try to make a guess that is slightly better than a coin flip. I actually remember seeing a deck from a data company boasting how their gender data was ~ 60% accurate.

Why are we worried about this? Contextual targeting is far more effective. As an example, targeting an ad on Elle.com is more than 60% likely to reach women than an ad targeted based on cookie data.

Genuinely curious so would appreciate calm, non-aggressive responses. I'm wondering whether this is due to a generally poor understanding of cookie targeting effectiveness or if I'm missing something.

Could you elaborate on why you believe security tokens to be a bad idea?

The only thing that stands out in that piece is that they're collecting advertising device IDs. Those IDs are anonymous and resettable. Snapchat doesn't need them internally since they have a more persistent identifier, i.e. snapchat user ID.

It might be that they want to pass this ID into programmatic platforms. Some systems treat them as pseudo-cookies for frequency capping, measurement, and basic targeting.

There’s likely strict quality control being enforced by Snapchat. It will be interesting to see them scale this.

Given that a small number of companies are responsible for a large proportion of brand advertising spend, I would posit that it isn't too difficult to scale quality control if you focus on building relationships exclusively with heavy spenders. Quality control becomes a bottleneck at the long tail end of the market, where ad budget per creative is much smaller. At the top end, you can behave like a TV channel and reuse established quality control practices. If you have 100m+ active users and a video-heavy medium, this shouldn't be impossible to accomplish.

That said, I have heard several reports about Snapchat's eagerness to build programmatic ad expertise, which (in the context of this discussion) may pointing to one of the following possibilities:

1. Large brands are not exhibiting TV-style buying behaviour (i.e. large upfront buys) on Snapchat. This would make sense as it mirrors trends in the rest of the industry. If you, as the buyer, have access to technology that lets you better control and manage your spend, gather intelligence, and understand ROI, then why wouldn't you? This is the buying mechanism that programmatic technology enables, and its more or less incompatible with TV-style upfront buys.

2. Snapchat wants to open their doors to mid-market and long tail advertisers. The goal might be to boost spending, drive up prices, or to go after performance advertising dollars. If I were a Snapchat investor, I'd find this second scenario slightly worrying, as quality control becomes very very difficult.

In this case, thoughts do seem to be the upper bound. If we are talking about linear streams of vocalised thought, then I suspect we might not get much further from where we are. I, for one, find it harder to stream thoughts faster than I can talk.

But consider any highly complex communication task. For instance, say you have to explain how recursion works. You have to go through abstract thinking, analogising, structuring, word selection, and delivery. It's hard work, and unless your execution is spot on, vital information maybe lost in transmission. Importantly, your thoughts here may not be a linear vocalised stream, and the process may not take very long.

It seems that tasks like these would benefit from some form of information exchange mechanism that doesn't rely on information being encoded in present day language. It would be most efficient to somehow "compress" your thought process around recursion as you're thinking about it, and then have the receiver unzip it on reception.

Yes, I'm wondering if there is any potential for leap frog advancements.

I see your point, and I would argue that both the rate of output and quality of output matter. You're right that one doesn't necessarily count without the other. A human could produce hundreds of gigabytes of pseudo-random numbers with a single key press, but that output could hardly be classified as 'information'. To classify any data as information, it has to represent some meaning that can be discerned by a receiver with at least some degree of fidelity. Humans do appear to be limited in our ability to produce information at high rates.

Honest question for people who are opposed to this update and think it should be opt-in: what do you see as the downsides or pitfalls of sending anonymised usage info to Homebrew / Google Analytics?

I think this is a well thought out list. I find it saddening that the 'health' of the web has deteriorated to a point where it's arguably uncompetitive for a vast number of modern applications including information discovery. I find it troubling that information discovery is moving away from the open web into non-neutral native application environments such as Facebook, Snapchat, and Twitter. Sure, information discovery on the web was driven by search engines. But I think search engines are really just part and parcel of the web (and arguably a feature that ought to have been built into web browsers from the beginning). A search engine doesn't express bias towards any information source (at least not by principle).

What troubles me is imagining a future where a vast majority of information that we consume is selected and curated by commercially driven, black box, non-neutral platforms.

I read an article recently about how Google is experimenting with removing the need to download apps by 'streaming' app content through a search box. Presumably, this is how Google search stays relevant in the post-web era where more and more information flows through walled gardens installed on mobile devices. Now, instead of re-inventing the web, shouldn't we be working towards 'fixing' the one that exists? This list provides a decent 10,000 ft overview of the problems.

Eager to hear your thoughts.

i'm sure the ad networks are laughing their way to the bank with an extension like this.

Definitely not. If this were to catch on, the whole ad ecosystem would feel the impact. Ad networks have advertiser relationships to look after. If the numbers looks fishy, they are held to answer by clients and usually instructed to stop running ads on sites where the suspicious activity originated.

Staying somewhere cheap gives you a tangible advantage but also puts you at several intangible disadvantages. At my first startup in the UK, we decided to base out of Birmingham instead of London, which reduced our living expenses by 70% and allowed us to really stretch our seed funding. But we ended up having to commute to London frequently for important meetings. Also, one could argue that we might have accomplished more in less time had we had continuous access to expertise and contacts at scale. Eager to hear other thoughts and experiences.

In its defense, the Economist has been known for its handsome catchphrases since before the modern clickbait era began. Take their covers for example. I don't think they're bowing to pressure -- just doing what they do.