AI Slows Entry-Level Hiring, Not Senior Jobs
ALSO: Gates’ $1B AI pledge, Salesforce debuts Koa
Welcome to Guru99 AI Report!
Top Story: If you’re early in your career, this one’s worth a read — AI is quietly reshaping who gets hired, and it isn’t seniors feeling the squeeze. We’ve also got Salesforce’s in-house AI, Google’s health breakthroughs, and a $1B pledge worth watching. Let’s dig in.
📉 AI’s Hiring Squeeze Is Real, But Narrow
Brief Buzz:
AI is now showing up in the job figures. According to a recent report from Goldman Sachs, it is slowing hiring — but only in a small number of highly exposed sectors, and mainly for entry-level workers. The broader job market is still holding its ground.
- In the US, call-centre employment is 39% below trend, in Canada 33% below, and in Germany 27% below.
- The impact hits entry-level workers hardest across more than 800 occupations — firms are hiring fewer juniors rather than laying off senior staff.
- AI adoption sits at 15–20% in developed economies (France, the US, the Netherlands and the UK lead) and 10–15% in emerging markets.
- Goldman calls the effect real but limited, touching only a few exposed industries and workers.
- A separate Ramp study found heavy AI users grew headcount about 10% and junior hiring 12% — a correlation, not proof.
💡 Why Should You Care?
If you’re just starting out or working in an AI-exposed field, keep a close eye on hiring trends — the pressure is genuine, but it stays narrow rather than widespread.
🌲 Salesforce’s New AI Model Koa Keeps Data In-House
Brief Buzz:
Salesforce has built its own AI. At Dreamforce 2026 it introduced Koa, a reasoning model built for CRM systems. The aim is to keep company data on-premises while AI agents handle multi-step business tasks — and Salesforce isn’t the only one doing it.
- Koa runs on Nvidia’s open Nemotron 3 Super and was fine-tuned on synthetic data — Salesforce says no customer data was used.
- Salesforce claims Koa matches or beats leading models on its own CRM benchmark, with just one-third the errors.
- It also introduced AIforce, which feeds CRM data out to tools like Claude (via Claudeforce) and Slack.
- Koa is in pilot now and becomes generally available this winter in US regions.
💡 Why Should You Care?
Own the AI layer and your company’s data — along with its advantage — stays in-house instead of going to a vendor that also serves your competitors. Expect more firms to follow.
🛑 OpenAI Asks Congress If Slowing AI Is Legal
Brief Buzz:
AI safety has turned into a legal question. OpenAI recently asked Congress whether rival labs could agree to slow advanced AI without breaking antitrust law — and it comes as fresh evidence shows AI systems both misbehaving on their own and being misused by people.
- A WIRED report says OpenAI asked lawmakers whether coordinating an industry-wide AI slowdown with rivals could breach the Sherman Antitrust Act.
- It follows OpenAI chief scientist Jakub Pachocki’s essay calling for “voluntary slowdowns” until shared safety standards exist.
- Turing Award winner Yoshua Bengio explains that AI agents learn to lie, cheat, and self-preserve because training rewards whatever gets results.
- Anthropic’s threat report details misuse: espionage linked to Russia, a missile-guidance cell in Yemen, and surveillance in Mali involving about 25 million SIM cards.
- In two weeks, a China-based network ran more than 4,700 AI dating profiles that messaged over 25,000 people.
💡 Why Should You Care?
Two problems now converge — what AI does on its own and what people make it do. Whether regulators allow safety “brakes” without also enabling collusion will set the pace of AI’s development.
🌍 Google Turns AI Toward Disease, Disasters, and Languages
Brief Buzz:
Google is making the case that AI can benefit humanity. In a recent blog post it said its tools now cover more than 300 languages and reach some 7 billion people, while highlighting fresh advances in genetics, weather, and crisis prediction.
- Google’s tools now support more than 300 languages, used by 86% of the world’s population — its biggest language-access milestone yet.
- The AlphaGenome Atlas is now open to researchers and covers all 9 billion possible single-letter DNA changes.
- WeatherNext 3, Google’s new weather model, delivers rain forecasts it says are 50% more accurate.
- The new Planetary Prediction Engine can forecast disease outbreaks, food shortages, and climate risks for emergency responders.
- AI is being applied to detect breast cancer, tuberculosis, and diabetes, while wearables flag heart and metabolic conditions.
💡 Why Should You Care?
This could mean earlier cancer detection, faster disaster warnings, and everyday tech in your own language — though Google itself cautions the benefits aren’t guaranteed and depend on wide collaboration.
🔬 Scientists Say AI Saves 7 Hours Weekly — With a Catch
Brief Buzz:
A new study from Google, DeepMind, and MIT FutureTech looked at how scientists actually use AI. Drawing on 15 million Gemini conversations, more than 2,600 specialized models, and a survey of 637 researchers, it finds real time savings — but new bottlenecks too.
- Scientists use AI more than most workers — nearly half of those surveyed say they use it every day.
- The work splits between general chatbots and specialized tools like AlphaFold: one writes and codes, the other predicts.
- Researchers say they save around seven hours a week, most of which they reinvest into more research.
- The catch: 41% say their pile of untested hypotheses is growing, because physical experiments can’t keep up.
- About half say AI nudges them toward safer, more incremental questions; only 28% say it makes them ask riskier, more ambitious ones.
💡 Why Should You Care?
AI is speeding up parts of research rather than replacing scientists — but because it generates more ideas, more need verifying, so the biggest breakthroughs may wait on slower physical experiments to catch up.
🤝 Gates Foundation Pledges $1B to Close the AI Gap
Brief Buzz:
The Gates Foundation has pledged one billion dollars to keep AI from benefiting the rich more than everyone else. The two-year commitment will put AI tools into health clinics, classrooms, and farms in the world’s poorest regions.
- One billion dollars over two years, drawn from the foundation’s core budget — a “down payment,” says CEO Mark Suzman.
- The split is 40% education, 40% health care, 10% agriculture, and 10% digital tools such as local-language datasets.
- The report notes today’s AI struggles with most languages — the error rate in Yoruba runs as high as 60%, versus under 6% in English.
- The 10th Goalkeepers report argues AI’s path isn’t fixed, but the window to shape it is closing fast.
💡 Why Should You Care?
Right now AI works best in English and for the wealthy — this is a bold early move aimed at changing that for everyone else.
Hey! I’m Krishna Rungta
Founder of Guru99.com, Editor-in-chief & Technology Expert
Was this email forwarded to you? Sign up for free here.
