Interim government's development spending falls short by 61%

Morning! You know what's super special about today? It's our 1st anniversary together 💐 February was the month we first launched Itla, and we don’t know how we can thank you for your love … actually, we do.

We're rolling out a new segment called Itla Careers: Where Info Meets Inspo 💡where each week, we'll sit down with industry pros for a quick Q&A—just 3 questions—to get their career insights and inspirations, whether they're based in Pakistan or abroad.

We can't wait to dive into these behind-the-scenes chats and bring you closer to the topics we often cover in the news. Our replies are always open, so feel free to share your feedback/suggestions 📩

In this week’s edition:

🛍️ What shopping will look like in a few years?

🏛️ Target missed by 61%

💬 Your new workplace bestie

👩‍💻 The lost ‘it’ factor of the IT industry

- The Itla Squad 💼

Itla Careers

Where Info Meets Inspo 💡

Where Info Meets Inspo is an Itla segment where we highlight professionals who work within the tech and business realm - in Pakistan or abroad. Hit reply if you’d like a chance to be featured.

How would you explain your job to someone with no exposure to your field?

Hi! I’m a Data Scientist in the Artificial Intelligence team at Afiniti, a company that works to maximize call center revenues by finding optimum pairs of agents and callers that would lead to a favorable outcome set by our clients. I would divide the work I do into 3 parts.

The first and most important is the modelling part, where we aim to create caller groups and approximate the performance of agents, using algorithms like decision trees and Bayesian mean regression. We run massive grid searches until we find the right hyper parameters for our model to deploy to production. You think the hard work would be over now, but a lot of time is spent in the next phase: production monitoring. This is an essential part of any job that works with real time events taking place (think bank transactions or food delivery). We want to ensure the model is performing as expected, and in case any anomalies arise, we must catch them and resolve them immediately. This can make for quite a fast/paced, and at times stressful, work environment and feels no less than an adrenaline rush! Thirdly, a good amount of time is spent on debugging when the model isn’t performing well. We conduct analysis, assessing the performance using various metrics which can lead to some actionable insights. Then we start working on a new model to deploy and the cycle repeats again :)

What’s the most interesting project you’ve worked on?

Since a lot of my work in the team is spent on training and deploying models, I have done quite a bit of work in automating and streamlining parts of the machine-learning pipeline. I would say it has been the most compelling since it involved really understanding how the pipeline works and testing it multiple times. Since our client had quite a complex set up, there were many additional parts to add. It helped me get really good at R and I also got to learn about MLOps. And after optimising some parts, getting the pipeline to run much more efficiently was very satisfying.

What’s your favorite tech/business-related media you read/watch/listen to?

I know that LLMs (Large Language Models) are the buzzword nowadays. So for anyone who is eager to dive deeper into how they work, I wanted to share that SuperDataScience recently released a podcast episode where they explain how transformers work in detail. For those who don’t know, transformers are the building blocks of the LLMs that we see today like GPT and BERT, and here, even though it’s a podcast, it does a really good job of breaking down how it all works. They are essentially walking through the paper “Attention is All You Need” which introduced transformers that ended up being a groundbreaking technology. There are visual aids and a transcript to follow along more easily. I also see many articles coming out that try to speculate whether chatGPT will take over our jobs, but I believe if people instead tried to understand how the technology really works, they would have a better idea of what is really happening and they wouldn’t fall for the fear-mongering as easily.

Business

Global 🌎️

What shopping will look like in a few years?

Retail has changed a lot recently, and leaders are thinking about what might happen next. Luckily, CNBC talked to some top executives and leaders in retail to get some insight.

The details: Looks like the retail leaders have a lot of thoughts.

• They seem to have mixed feelings about the pandemic. COVID messed up retail big time. Ports got jammed, stores ran out of stuff, then they had too much, and people started wanting different things because prices went up. Because of all the COVID chaos, some things that were slowly happening in retail sped up. Like more people doing pickup and using apps to shop.

• Retail bosses think stores will become smaller and more personalized. They want stores in places where customers want them, not just anywhere.

• Leaders believe technology like AI and data analysis will make a big difference. It can help predict what customers want and make shopping better.

• People will still shop in stores a lot, but they'll decide what to buy online first. Retailers need to be good at both online and in-store shopping to keep people happy.

• The future of retail is about making shopping more fun and interesting. Stores will be like adventures, mixing real and digital experiences.

TLDR: Retail is evolving fast, and CNBC spoke with top executives about what's coming next. Leaders foresee smaller, personalized stores and AI transforming the shopping experience. Stores will still be crucial, but more interactive and adventurous, blending real and digital experiences.

What else is new in news

✶ You might wanna move to Texas if you want to be a tech-bro in a big firm. A new report from MIT found that poorer counties with lower employment rates are receiving a disproportionate share of private sector investment in strategic sectors compared to their economic output and population.

Local 🇵🇰

Government development spending falls short by 61%

The Ministry of Planning reported that government spent Rs191 billion on development projects in the first seven months of the fiscal year. However, this amount is much lower than the target of Rs494 billion, showing a significant shortfall of 61%. This shows there's not enough money going towards things like building roads or improving infrastructure.

The details:

• To meet certain financial targets set by the IMF, the government slowed down how much money they were spending. Even though they tried to spend less, they still didn't hit their targets. This means there were problems with how they released funds and how quickly they completed projects.

• Pakistan promised the IMF they would speed up spending and save money on certain projects. They also decided to transfer some projects to provinces and stop spending on certain things to save money.

• While the federal government didn't spend much on projects, provinces spent more than expected, especially Punjab. This shows that provinces are doing more to improve things like roads and other infrastructure compared to the federal government.

TLDR: The government spent Rs191 billion on development projects, missing the Rs494 billion target by 61%. Despite attempts to cut spending, they fell short, leading to issues with fund allocation and project completion. Pakistan pledged to accelerate spending and transfer projects to provinces. Provinces like Punjab exceeded expectations, focusing more on infrastructure improvement.

News Flash

✶Even with sky-high interest rates, crazy auto financing, and all those high prices, along with the usual political and economic rollercoaster, the total sales of cars, jeeps, vans, and pickups increased by 81% in January. That's a jump from 5,816 in December 2023 to 10,536.

Tech

Global 🌎️

Your workplace BFF might have some competition

Now, slack with the help of AI will be like your workplace bestie and fill you in on all the office happenings when you get back from vacation (although, for office-gossip, you might have to rely on your human-workplace-bestie for now).

What happened: Slack, a popular workplace communication platform, has introduced the Slack AI Suite, a set of smart features designed to make work easier for Slack Enterprise users. This new suite uses advanced AI technology to help users stay updated, collaborate better, and find information quickly.

What Slack AI includes:

• Now, users can get quick summaries of Slack threads and channel conversations right within the platform like highlighting key points and who said what.

• You have the power to customize your summaries based on what matters most to you. Choose to see summaries of unread messages, recent messages from the past week, or messages within a specific time frame.

• Whether you need clarification on a project, workplace policy, or acronym, simply ask in Slack, and the AI will find relevant messages to provide you with the answers you need, making your workflow smoother.

• Slack AI seamlessly works with other apps within Slack, such as Notion and Box. This means you can quickly access summaries of documents without leaving Slack.

TLDR: Slack's new AI Suite brings smart features to Slack Enterprise users, helping them stay updated and collaborate better without leaving the platform. It offers quick summaries of threads and channel conversations, customizable to your preferences. You can also ask questions and get relevant answers, plus integrate with apps like Notion and Box for easy access to documents.

What else is new in news

✶The dating app market isn’t as hot as before. A recent analysis by data.ai found that global downloads of dating apps grew only by 1.9% this January compared to last year. In the U.S., growth slowed too, with only a 2.38% increase in installs, down from nearly 16% in January 2023. 

Local 🇵🇰 

The lost ‘it’ factor of IT industry

Pakistan's IT export industry has a big chance to grow and help the country financially without depending too much on outside help. But there are some problems we need to solve first accroding to Data Darbar.

The details: Last year, Pakistan's official IT exports were worth $2.6 billion. Experts think these exports could grow by 20% every year, reaching $6.5 billion in five years and $16.1 billion in ten years.

• Many IT companies don't fully bring back the money they earn from abroad. There are reasons for this, like it's hard to change foreign money to Pakistani money, worries about higher taxes, and complicated rules for bringing money back.

• Around $2.7 billion is missing from the official records. Data Darbar looked at things like how many people work in the IT industry and how much money each person makes.

• Sometimes, companies keep their money in other countries. There have been times when the amount of money we earned from IT exports went down, even though the industry kept growing.

Zooming out: To fix this, we need to make it easier for companies to bring back their money and report it properly. By doing this and helping the IT industry keep growing, Pakistan could reach the goal of $19.3 billion in IT exports by 2033.

TLDR: Pakistan's IT export industry has huge potential to boost the country's economy significantly, aiming to not rely heavily on external aid. Despite facing challenges like money repatriation issues, high taxes, and complex regulations, experts believe with proper reforms, Pakistan could increase its IT exports from $2.6 billion to $19.3 billion by 2033.

Tech tour of news

✶Careem stated that a recent social media post, which drew criticism from the PML-N, was not related to the country's political situation. The post, featuring the words "Program war gaya?" (meaning "program gone bust?"), has since been removed from the platform.

More interesting Itla (اطلاع) we consumed:

🏴‍☠️ Reading this essay feels like reading a thriller movie script. You might think, "I would never do something like this," but as you read, you might start to feel really surprised and a bit sad. Charlotte Cowles wrote a story about how she got tricked into putting $50,000 in a shoebox and giving it to someone she didn't know in a big car.

It starts with a phone call from someone pretending to work for Amazon, saying something was wrong with her account. Then, she gets passed around on the phone to people who say they are from the government, telling her a scary story about someone stealing her identity. They make her so scared and confused that she ends up giving them almost all the money she has saved. It's a story that's hard to believe, but it shows how anyone can be tricked if the scammers are convincing enough.