hocuspocus

Pretty weak argument as most points aren't inherent to gradual typing at all.

GPT-5.4 mini seems noticeably better to me, token cost between Gemini 3 and 3.5 Flash.

I'm not sure about broadband data, as it can't be that useful. However on the mobile side, it's fairly valuable as a mobile app can collect A-GPS location and sensor telemetry that are unknown to the MNO otherwise.

And Javascript and Python the best languages.

I think the AI/ML ecosystem is a bit of a mess overall, things tend to work out of the box in Python because that's what everyone targets, but it doesn't necessarily say that much about maturity and robustness.

In Rust you can use many C++ frameworks like libtorch or ONNX or specialized libraries (llama.cpp, whisper.cpp ...) via their bindings. Native projects such as Candle or Burn are not feature complete yet, but I assume they'll eventually get there and drive bigger communities compared to C++.

Hu, no?

The most common scams nowadays involve social engineering to make people log into their online banking and transfer money, specifically because there's no way to revert transactions. The victims are typically 100% liable as they accept and authorize these.

https://www.youtube.com/watch?v=VtbihSwrKhk

https://www.srf.ch/sendungen/kassensturz-espresso/kassenstur...

Which has zero incentive to side with the other party. You pay by bank transfer, once the money is on the merchant's account, why would their bank ever agree to a refund in case of dispute? Now it's between you and the merchant, good luck with filing police reports and court claims especially abroad.

If EPI / Wero reach a critical mass then hopefully this will change, as non-Eurozone countries will have a direct interest in making the local mobile payment solution accept the one used by an area of 350+ million people.

Then the EPI protocol could become the least common denominator and you might be able to use Vipps, Twint, Blik... in Sweden? I believe similar scenarios are happening in Asia around Alipay and UPI, for instance I think I can use the Korean Kakao Pay on a payment terminal in Japan, because both sides are compatible with Alipay+.

This is a false dichotomy. You can have very cheap transaction fees, Pix is state-run and probably operates at a loss, with merchant fees as low as 0.2~0.3%. In comparison the cheapest card payment under the EEA interchange cap is probably slightly above 0.5% when you add scheme fees and PSP costs.

However businesses do require payment systems and not just barebones bank transfers. Except for high trust, low volume transactions such as buying a car, paying your rent...

The UK has had instant transfers for many years.

Switzerland since last summer, but banks aren't forced to expose the feature to retail customers yet, at least not for free. We could have had that 15 years ago if our government wasn't so afraid of upsetting the banking lobbies.

> We need instant, free SEPA transfers around the clock.

We have? SCT inst has been rolled out almost everywhere

> Switzerland is not part of SEPA

We definitely are. Transfers in EUR from/to CH IBANs use SEPA rails. CHF accounts can also send or receive EUR transparently (usually at a bad FX rate, but it just works).

> I don't understand what they solve. [...]

> I always say that the day Trump decides to block Visa/MasterCard outside the US is the day we get instant payments and finally get rid of cards.

You answered your own question. We need pan-European payments systems on top of the existing banking infrastructure. Payments and transfers aren't the same thing. By moving to mobile wallets with QR code and NFC payments, this opens up interoperability beyond Europe too.

Swish is even older.

The issues with Twint:

- Switzerland hasn't gotten Apple to open up NFC payments at the same conditions as the EEE, at least not yet.

- Twint is part of the EMPSA which hasn't really delivered anything tangible. On the contrary there's now a real push from EPI and EuroPA to make Wero interoperable with non-Eurozone networks. Hopefully Twint will get on board too.

- Transaction fees are on the higher side, consequently merchants don't have any reason to push Twint or disincentive debit card payments.

Even better, as an exchange, they don't even necessarily care whether the line goes up, down, sideways, or in fucking circles to quote the Wolf of Wall Street. As long as it goes somewhere, and customers are charged fees.

PANs are indeed going away and every transaction could already be tokenized, today. But then the US were 20 years behind on EMV, and SCA is still not a thing.

The EU has banned plain SMS tokens for SCA. You need an OTP + PIN or password, or more likely authorize the transaction from a mobile app with biometrics.

I work at a carrier in Switzerland, if I take the phone we push at the top of our inventory, the iPhone 17 Pro 256 GB:

Phone: 1099 (Apple's MSRP). 24-month phone plan, 50/month the first year, then 81/month: 1572. Discount: -300

Now BYOD:

Phone: 1067 at the biggest and most reputable online retailer. Near identical phone plan sold through our discount/virtual brand: 30/month i.e. 720 over 24 months (but you can leave anytime).

So even assuming you buy it on day 1 with zero discount from Apple's MSRP, that 300 gift costs you 852 CHF really.

Now if I look at our most agressive competitor, you get the phone for free if you spend at least 1752 on a 24-month phone plan.

They have a better phone plan at 15/month i.e. 360 over 2 years. BYOD is still way cheaper.

Carrier deals aren't comparable as they're subsidized by inflated phone plans.

If you buy an iPhone on day 1 from the Apple store you're certainly not getting any gift or discount. With Google you can pre-order and get a $200-300 gift, wait for deals (Black Friday, Christmas, ...) or if you aren't in a rush, wait six months for resellers and carriers to invariably start dumping their stock at half the MSRP. Meanwhile the base iPhone price has decreased by 10-15% maybe.

And I don't see how older MacBook Airs are relevant here. Apple has always sold previous generations of their hardware for years, directly or indirectly, either by actively maintaining production or simply letting resellers deal with old stocks. Google keeping the previous year generation on their product line-up is a very recent development. Not very long ago they would abruptly stop selling the Pixel N before the N+1 were even announced.

Google just sucks at this.

The Nexus and early Pixel eras have been a series of weirdly positioned and priced models year after year. Google didn't even bother trying to sell them in more than a handful countries.

Then Google switched to a pretty bad Exynos based SoC and consistently shipped underwhelming hardware at price points aligned with Apple... requiring absolutely insane discounts or promotion campaigns, even at launch. Can you imagine Apple giving you AirPods Pro or a Watch if you buy the next iPhone 18 on day 1?

We've seen things like the Pixel tablet that should have been sold at half its MSRP to stand even a small chance against the iPad.

And I won't even talk of the many hardware issues... my dad's Pixel 7a was fully reimbursed for 25% more money than he actually spent buying it.

Most monadic effects aren't executing side effects.

> Copilot revenue goes to 0 if GitHub burns to the ground.

That's not remotely true. I doubt most Copilot Business/Enterprise subscribers care about GitHub at all.

> 50M+ subs operator, at least 10 employees can have both location and CRM data, I guess it's pretty typical.

This shouldn't be the case anywhere in Europe or regions with similar laws. And we have a lot less than 50M subscribers.

Anyway, there's really nothing that justifies having access to both. If you work on network quality and need enriched traces, personal data is completely useless. Most business cases don't even need stable, let alone clear IMSI. Very few people will need to look at a clear MSISDN for troubleshooting, and if you do things properly they shouldn't get blanket access to terabytes of daily telemetry.

Aggregated CRM data can be useful to more high-level business cases, nothing that can be used to identify someone personally. Our data governance office doesn't even let us correlate anonymized and GDPR compliant data that we buy from third parties when the IDs are too stable, as it'd be fairly easy to match raw network traces.

> so you do have access :)

No I don't. Sometimes people move to different teams you know, and access to datasets I had in the past is mutually exclusive with some that I do have now.

> correct for LI, not for emergency.

If people that can see E112 payloads with GNSS locations exist, then I don't know they are, but I'm sure they can't have access to stuff relevant to the discussion here. On the network telemetry side, our job is monitoring and quality assurance. Anyway this kind of data is too sparse to be abused by a stalker.

So now the random ex-boyfriend has access to the same tools as 3 letter agencies, got it.

If you live in a country where you cannot trust law enforcement then there isn't much your telco can do. But specifically, these surveillance tools are not available to us.

Built-in positioning of network traces is relatively recent in mobile network equipment and dedicated probes.

If that happened more than 5-6 years ago, it would sound even less likely. Most telcos never bothered doing the processing needed to position raw events based on timing advances. They'd simply offload that to third party companies. These solution providers aren't crazy, they don't touch data that isn't already anonymized. It's even less probable that a random employee would have access to the multiple datasets needed to piece someone's personal data together.

I'm specifically talking about the technical aspect. Even with non-existent separation of concerns, and abysmal practices related to data governance which would be breaking the law in most of the developed world, the story sounds like bullshit. Extracting points of interest and reconstructing paths from raw network telemetry isn't trivial.

The likelihood a random employee could run a quick SQL join to stalk someone based on their name is zero.

As far as telcos go, I work at a pretty small one. We have fewer subscribers than say, a single Chinese operator would have in a second tier city.

Even in pretty dysfunctional countries, or pro-business ones like the US, where nothing like the GDPR exists, telcos management have a strong interest in not letting just any rank and file employee spy on subscribers.

Absolutely not. I have access to geo-located network telemetry. CRM data is completely off limit to anyone on my team.

I'm sorry but this sounds like bullshit. As someone who has access to such data at a telco:

- Very few people have legit business cases requiring access to enriched network telemetry, at least non aggregated.

- Of which, only a handful have any reason to see the MSISDN in clear.

- Of which, none can get access to clear CRM data.

- Lawful interception and emergency services use completely separate paths, exposed via user interfaces that aren't available to employees.

And obviously, a simple email to the data governance and privacy office would be taken extremely seriously.

Also why not simply switch to a different phone operator?

It is vibe-coded by people from Vates, the company maintaining https://github.com/xcp-ng

Their internal IT infrastructure runs self-hosted OSS wherever possible. I don't think cal.rs is a toy project, they know the perils and headaches of doing open source.

I think it's fine. GitHub Copilot is popular as ever, especially in companies that have enterprise tier subscriptions. Plans for personal use pretty good too, pricing is competitive. The VS Code integration and agentic features aren't bad either.

Developer tools live in their own space. And I assume most devs don't really care that "Copilot" started to show up everywhere, especially in MS365 products. At least I don't. Conversely, do non-technical people care where the term comes from, and now means "LLM integration" in a bunch of MS products?

I think it's better that Google going through Bard, Gemini, IDX, Firebase Studio, Antigravity, ...