180999726
submission
ZipNada writes:
Over the weekend, the OpenAI cofounder and former director of AI at Tesla posted a graphic showing how susceptible every occupation is to Al and automation, using Bureau of Labor Statistics data. Different jobs received scores on a scale of 0 to 10, with 10 being most exposed.
While the overall weighted exposure was 4.9, Karpathy’s data also showed that professions earning more than $100,000 a year had the worst average score (6.7), while the those earning less than $35,000 had the lowest exposure (3.4).
180839462
submission
ZipNada writes:
Gopal said that arguably, many businesses exist that AI can never be trained on, “because this is real-life business that moves.” Real people who have conversations and continually update a business context will always be one step ahead of the machines, he explained. “Are you going to retrain for that one individual conversation for one day?” he asked, and then retrain on a rolling basis every time your business context changes?
Gopal was bearish about how much this context can be captured, estimating that 70% of the effort required to make AI useful relies entirely on unwritten business context that exists only in human heads. “You fundamentally cannot train a system” on this fluid daily reality, Gopal explained, noting that real-life business constantly changes based on individual conversations and human interactions. While AI can automate tasks at the absolute top (coding) and the absolute bottom (physical robotics), the vast middle ground of knowledge work requires human context.
180771790
submission
ZipNada writes:
Let me give you an example so you can understand what this actually looks like in practice. I’ll tell the AI: “I want to build this app. Here’s what it should do, here’s roughly what it should look like. Figure out the user flow, the design, all of it.” And it does. It writes tens of thousands of lines of code. Then, and this is the part that would have been unthinkable a year ago, it opens the app itself. It clicks through the buttons. It tests the features. It uses the app the way a person would. If it doesn’t like how something looks or feels, it goes back and changes it, on its own. It iterates, like a developer would, fixing and refining until it’s satisfied. Only once it has decided the app meets its own standards does it come back to me and say: “It’s ready for you to test.” And when I test it, it’s usually perfect.
I’m not exaggerating. That is what my Monday looked like this week.
I’ve always been early to adopt AI tools. But the last few months have shocked me. These new AI models aren’t incremental improvements. This is a different thing entirely.
The experience that tech workers have had over the past year, of watching AI go from “helpful tool” to “does my job better than I do”, is the experience everyone else is about to have. Law, finance, medicine, accounting, consulting, writing, design, analysis, customer service. Not in 10ten years. The people building these systems say one to five years. Some say less. The market was spooked enough this month that it wiped out $1 trillion worth of software value in just a week. And given what I’ve seen in just the last couple of months, I see more disruption, and soon.
180754968
submission
ZipNada writes:
AI Swarms Beat Traditional Teams on Speed, Cost & Startup Readiness
11:45 am February 4, 2026 By Julian Horsey
What if you could deploy an army of 100 AI agents to tackle your most complex projects in minutes, and at a fraction of the cost of traditional systems? Universe of AI walks through how the Kimi K2.5 Agent Swarm is redefining what’s possible in AI task management, offering a solution that’s not only 8x cheaper than Claude Opus 4.5 but also up to 4.5 times faster. Imagine launching a startup with a fully functional website, a complete marketing strategy, and detailed competitor analysis, all generated in under 10 minutes. This isn’t just a futuristic concept; it’s a reality that’s reshaping how businesses approach productivity and efficiency.
180391145
submission
ZipNada writes:
The explosion in AI models, software, and agents has raised questions about the impact of the technology on the broader job market as companies find new efficiencies from this new technology.
But according to EY's latest US AI Pulse Survey, just 17% of 500 business executives at US companies that saw productivity gains via AI turned around and cut jobs.
"There's a narrative that we hear quite frequently about companies looking to take that benefit that they're seeing and put it into the financial statements reducing costs, or cutting heads," EY global consulting AI leader Dan Diasio told Yahoo Finance.
"But the data that we asked those 500 executives does not bear that out. That is happening less than one out of five times, and more often they are reinvesting that," he added.
180196635
submission
ZipNada writes:
A new generation of stratospheric balloons and high-altitude uncrewed aerial vehicles (UAVs) could soon connect the world's unconnected with high-speed internet at a fraction of the prices commanded by operators of satellite megaconstellations such as SpaceX's Starlink.
High-altitude platform stations, or HAPS, have been around for a while, but the technology hasn't fully taken off yet. Google spent 10 years trying to develop balloons that would hover in the stratosphere above remote rural areas and beam internet to residents but abandoned that project, called Loon, in 2021, concluding that it couldn't be made sustainable.
Four years later, companies such as World Mobile Stratospheric and Sceye say they are on the verge of making internet-beaming from the stratosphere, the layer of Earth's atmosphere roughly 6 miles to 31 miles (10 to 50 kilometers) above the planet, a reality. Moreover, they claim that their offerings will be better and cheaper than that of satellite megaconstellations in low Earth orbit (LEO), which too have been developed with the promise of connecting the world's unconnected.
178979670
submission
ZipNada writes:
More than 85 veteran climate scientists have pushed back against a Trump administration report downplaying the severity of climate change, submitting more than 400 pages in public comments to the Energy Department on Tuesday.
The department’s Climate Working Group report, released July 29 alongside proposals to deregulate some polluting sectors, was authored by five well-known climate change contrarians and even portrayed climate change as potentially beneficial.
178122443
submission
ZipNada writes:
The essence of search is compression: distilling insights from a vast corpus. Subagents facilitate compression by operating in parallel with their own context windows, exploring different aspects of the question simultaneously before condensing the most important tokens for the lead research agent. Each subagent also provides separation of concerns—distinct tools, prompts, and exploration trajectories—which reduces path dependency and enables thorough, independent investigations. ...
these architectures burn through tokens fast. In our data, agents typically use about 4× more tokens than chat interactions, and multi-agent systems use about 15× more tokens than chats. For economic viability, multi-agent systems require tasks where the value of the task is high enough to pay for the increased performance. Further, some domains that require all agents to share the same context or involve many dependencies between agents are not a good fit for multi-agent systems today. For instance, most coding tasks involve fewer truly parallelizable tasks than research, and LLM agents are not yet great at coordinating and delegating to other agents in real time. We’ve found that multi-agent systems excel at valuable tasks that involve heavy parallelization, information that exceeds single context windows, and interfacing with numerous complex tools.
177126677
submission
ZipNada writes:
Microsoft's latest research identifies a new type of organization known as the Frontier Firm, where on-demand intelligence requirements are managed by hybrid teams of AI agents and humans.
The report identified real productivity gains from implementing AI into organizations, with one of the biggest being filling the capacity gap — as many as 80% of the global workforce, both employees and leaders, report having too much work to do, but not enough time or energy to do it. ...
According to the report, business leaders need to separate knowledge workers from knowledge work, acknowledging that humans who can complete higher-level tasks, such as creativity and judgment, should not be stuck answering emails. Rather, in the same way working professionals say they send emails or create pivot tables, soon they will be able to say they create and manage agents — and Frontier Firms are showing the potential possibilities of this approach. ...
"Everyone will need to manage agents," said Cambron. "I think it's exciting to me to think that, you know, with agents, every early-career person will be able to experience management from day one, from their first job."
176652755
submission
ZipNada writes:
The ubiquitous ESP32 microchip made by Chinese manufacturer Espressif and used by over 1 billion units as of 2023 contains an undocumented "backdoor" that could be leveraged for attacks.
The undocumented commands allow spoofing of trusted devices, unauthorized data access, pivoting to other devices on the network, and potentially establishing long-term persistence.
This was discovered by Spanish researchers Miguel Tarascó Acuña and Antonio Vázquez Blanco of Tarlogic Security, who presented their findings yesterday at RootedCON in Madrid.
"Tarlogic Security has detected a backdoor in the ESP32, a microcontroller that enables WiFi and Bluetooth connection and is present in millions of mass-market IoT devices," reads a Tarlogic announcement shared with BleepingComputer.
"Exploitation of this backdoor would allow hostile actors to conduct impersonation attacks and permanently infect sensitive devices such as mobile phones, computers, smart locks or medical equipment by bypassing code audit controls."
The researchers warned that ESP32 is one of the world's most widely used chips for Wi-Fi + Bluetooth connectivity in IoT (Internet of Things) devices, so the risk of any backdoor in them is significant.
176605441
submission
ZipNada writes:
Environmental Protection Agency chief Lee Zeldin has privately pushed the White House for a rewrite of the agency’s finding that planet-warming greenhouse gases put the public in danger. The original 52-page decision in 2009 is used to justify and apply regulations and decisions on heat-trapping emissions of greenhouse gases, such as carbon dioxide and methane, from the burning of coal, oil and natural gas.
“Carbon dioxide is the very essence of a dangerous air pollutant. The health evidence was overwhelming back in 2009 when EPA reached its endangerment finding, and that evidence has only grown since then,” said University of Washington public health professor Dr. Howard Frumkin, who headed the National Center for Environmental Health at the time. “CO2 pollution is driving catastrophic heat waves and storms, infectious disease spread, mental distress, and numerous other causes of human suffering and preventable death.”
That 2009 science-based assessment cited climate change harming air quality, food production, forests, water quality and supplies, sea level rise, energy issues, basic infrastructure, homes and wildlife. ...
It’s these indirect effects on human health that are “far-reaching, comprehensive and devastating,” said Katharine Hayhoe, an atmospheric scientist at Texas Tech and chief scientist at The Nature Conservancy. She said rising carbon dioxide levels in the air even “ affect our ability to think and process information.”
Scientists said the Trump administration will be hard-pressed to find scientific justification — or legitimate scientists — to show how greenhouse gases are not a threat to people.
“This one of those cases where they can’t contest the science and they’re going to have a legal way around,” Princeton University climate scientist Michael Oppenheimer said.
176189169
submission
ZipNada writes:
So let’s address the thorny issue of what agentic AI actually is: taking processes and tasks that a person would normally need to perform, where decisions would need to be made by a human on the outputs that preclude automation, and involving GenAI plus RPA and automation to fulfill it. Different agents specialize in doing different tasks, with some focused on compliance and standards. Some fulfill user requests; some seek out, collect and redistribute data to the right places. Workflow agents identify APIs as well as generate and execute workflows across applications.
As an example, in our company's world of QA testing, when tests are running and going through a new version of the application, there are AI agents running autonomously, making decisions. If the code behind the "button" has changed, these agents will be able to make real-time decisions on whether to fix this and keep running or stop the process. This is real-time diagnostics and fixing.
No one’s cowering before their robot overlords, though. Humans are still 100% going to be needed, but their roles will change and be less siloed. There will be fewer of them, too. One action from a team will have the potential to affect other teams much more quickly, making supervision of AI agents, roles involving training AI models and human oversight of critical decision making absolutely vital.
176104107
submission
ZipNada writes:
With ChatGPT and their ilk—Anthropic’s Claude, for example—you can find use cases, but it’s hard to point to any of them that are really killer apps. It’s impossible to point to anything that justifies the ruinous financial cost, massive environmental damage, theft from millions of people, and stealing of the entire internet. Also, on a very simple level, what’s cool about this? What is the thing that really matters here?
176068967
submission
ZipNada writes:
In results that came out late last year, the average scores of Americans ages 16 to 65 fell in both literacy and numeracy on the globally administered Program for the International Assessment of Adult Competencies.
And even among American adults, achievement gaps are growing. The exam’s results are broken down into six performance levels. On the numeracy portion, for example, the share of Americans scoring at the two highest levels rose two points, from 10% to 12%, while the percentage of those at the bottom two levels rose from 29% to 34%. In literacy, the percentage of Americans scoring at the top two levels fell from 14% to 13%, while the lowest two levels rose from 19% to 28%.
175870193
submission
ZipNada writes:
You are wearing AI-powered glasses that can see what you see, hear what you hear, and process your surroundings through a multimodal large language model. Now when you spot that store across the street, you simply whisper to yourself, “I wonder when it opens?” and a voice will instantly ring back into your ears, “10:30 a.m.”
By 2030, we will not need to whisper to the AI agents traveling with us through our lives. Instead, you will be able to simply mouth the words, and the AI will know what you are saying by reading your lips and detecting activation signals from your muscles.
By 2035, you may not even need to mouth the words. That’s because the AI will learn to interpret the signals in our muscles with such subtlety and precision — we will simply need to think about mouthing the words to convey our intent. You will be able to focus your attention on any item or activity in your world and think something and useful information will ring back
