A few years ago I worked for a company that no longer exists today. They had, among other things, a job search service connected to their ID system. I was also doing my own project at the same time and needed Python developers. I was young and naive and thought I could find a junior and train him quickly. So I posted a vacancy on this job board: "Looking for a Python developer without experience." It turned out that they showed my phone number and it couldn't be turned off.
I received about 3-4 calls from very strange people who demanded to know how to become programmers. For some reason, they all started calling at about 5am. I even gave some useful advice to the first one, because I was taken aback by such impudence.
Today I use about 4 different phone numbers to separate my private life and data leaks like that.
My wife cold-called about 60% of all businesses in our niche within our city. Meanwhile, I literally walked into CEOs' offices, asked politely for meetings, and pitched face-to-face. Conversion was around 1%, but that gave us our first collision with reality—and our first paying customers.
Also half of my clients came from Google search ads. But it was absolutely terrible - 9 out of 10 requests were people trying to talk to chatGPT
> But it was absolutely terrible - 9 out of 10 requests were people trying to talk to chatGPT
I don't totally follow, probably because I don't tend to run ad campaigns and I'm not up with whatever shenanigans advertisers are doing these days. What do you mean when you say that they were "trying to talk to chatGPT"?
Indeed, it was written in a rather unclear way. I have AI in the name of my company and I left a form on the site to fill out.
My site, due to a cold start, began to appear in the search results for people from different countries who misspelled "ChatGPT for free" and stuff like that. They went to the site, saw my form, filled out the required fields and wrote "Make me an essay on the topic...". No stop words can cover such behavior.
Importantly, all real clients did not use this form, but immediately went through convenient communication channels.
What do you mean by "cold start"? I can't figure it out, but are you saying because your site was very new google gave it more weight for those search terms and that's why you were ranking well on that competitive search terms?
Quite the opposite. A new site that Google is afraid to show for any normal queries. There is too much competition there. But there is a lot of space in the unique garbage results.
My guess is that the GP put a AI chat bot on their site, and people started asking it for things completely unrelated to his business as a way to get free access to chatgpt without an account.
It’s actually quite fun, whenever you see an ai chat bot asking you if you have any questions ask it for its favorite cupcake recipe. If it answers its a poorly configured LLM.
Chatbots went from costing $100K+ and months of development to $100 and a weekend—just in 5 years.
Here's how the rise of LLMs, cheap APIs, and simplified tooling transformed chatbot creation, why building custom solutions today is usually a bad idea, and what it means for the market going forward.
But this is an image-like benchmark. Has anyone looked at the article about the EU-ARC, what is the difference? Why can't you measure it on a regular one?
I glanced through it, didn't find it right away, but judging by their tokenizer, they are learning from scratch. In general, I don't like this approach for the task at hand. For large languages, there are already good models that they don't want to compare with. And for low-resource languages, it is very important to take more languages from this language group, which are not necessarily part of the EU
Why would they want more languages from outside of the EU when they've clearly stated they only target the 24 official languages of the European Union?
For example: Slovene language. You simply don't have enough data on it. But if you add all the data that is available on related languages, you will get a higher quality. LLM fails with this property for low-resource languages.
I'm not sure I'm convinced. I speak a small European language and the general experience is that LLMs are often wrong exactly because they think they can just borrow from a related language. The result is even worse and often makes no sense whatsoever. In other words, as far as translations go, confidently incorrect is not useful.
I'd love for everything described in this article to be true. If the U.S. really were pushing talent away en masse, maybe my odds of getting in would go up. But I have serious doubts.
Very few folks are rushing to pay half their salary in taxes, wrestle with EU bureaucracy, and earn 2-3x less pre-tax just to move to Europe. And don’t get me started on bottle caps screwed on so tight you can't drink from them.
I could go on about memeI'd love for everything described in this article to be true. If the U.S. really were pushing talent away en masse, maybe my odds of getting in would go up. But I have serious doubts.
Very few folks are rushing to pay half their salary in taxes, wrestle with EU bureaucracy, and earn 2–3x less pre-tax just to move to Europe. And don’t get me started on bottle caps screwed on so tight you can't drink from them.
I could go on about meme-tier Canada or even more meme-tier Australia. As for India or China - most people didn’t leave just to go back.
Sure, Trump is scaring off a few people, but most of those at risk - without roots, degrees, or real work experience - will fight tooth and nail to stay. Deporting one unlucky grad student probably won’t shift the odds in my favor. Pity.
Disclaimer: I’m IT-focused. No clue what’s going on with nuclear engineers or scientists. Maybe they live in a parallel universe.tier Canada or even more meme-tier Australia. As for India or China - most people didn’t leave just to go back.
Sure, Trump is scaring off a few people, but most of those at risk - without roots, degrees, or real work experience - will fight tooth and nail to stay. Deporting one unlucky grad student probably won’t shift the odds in my favor. Pity.
Disclaimer: I’m IT-focused. No clue what’s going on with nuclear engineers or scientists. Maybe they live in a parallel universe.
Has anyone looked at similar studies from, say, five years ago? I worry that half of what was labeled “delusional” back then is now a widely accepted opinion.
Feels like we're one step away from a follow-up paper titled "We regret to inform you: The paranoid were just early."
Not saying everything turns out to be true - just that social consensus tends to shift a lot faster than clinical definitions.
The clinical definition of delusion includes social consensus. From DSM-5:
"Delusions are deemed bizarre if they are clearly implausible and not understandable to same-culture peers and do not derive from ordinary life experiences. […] The distinction between a delusion and a strongly held idea is sometimes difficult to make and depends in part on the degree of conviction with which the belief is held despite clear or reasonable contradictory evidence regarding its veracity."
Probably the latter. Homosexuality was removed from the DSM as a diagnosis in 1973, by a vote of 5,854 to 3,810. Gender identity disorder was removed from the DSM in 2017.
It seems like there is a continuum between being oblivious and delusional. You need to notice things, recognize their meaning, make connections and then make conclusions. Being oblivious can mean missing important signs and making bad choices due to not realizing what is happening. Being delusional can mean jumping to conclusions by "connecting the dots" in one particular way that leads to one conclusion even though that may not be the truth. A sane perspective would mean evaluating different hypotheses to come to a reasonable conclusion. Someone who is paranoid may be right about some things much earlier than others, but they may also be wrong about many other things.
I think in many circumstances certain outcomes emerged because the paranoid willed it into existence. I get the feeling there are a lot doomsayers who secretly want the end of civilization to occur so that they can finally act out their fantasies, delusions, say they were right, etc.
Not sure about CBP specifically, but many countries already use specialized tools to break into phones and silently install backdoors — Cellebrite comes to mind.
As my favorite blogger puts it: "If the data is important, it's not stored in only one place. If there's no backup, it wasn't important." With that in mind, wiping the device and filling the gallery with high-resolution images of genitals covered in excrement remains one of the more effective passive defense strategies.
Jokes aside, it's depressing that crossing borders often means giving up fundamental digital privacy — and that we've largely normalized this. The idea that any government agent can dig through your phone without a warrant just because you're crossing a line on a map is dystopian at best.
This (along with healthcare) are examples for the "both parties are the same" refrain.
Neither side wants or advocates for the status quo (for US citizens to lose their rights at the border) but neither side is doing anything about it. They could easily eliminate the "constitution-free zone" exemption at borders and airports, but from what I can see, no lawmakers are talking about it.
If anyone starts, it's pretty easy to trot out some story about how CBP caught a drug smuggler or human trafficker or terrorist because of the searches and then the matter is quickly silenced.
I’m building an AI chatbot for businesses to handle sales and lead qualification via WhatsApp and Telegram — https://ziriusai.xyz/en
Doing it completely solo — built the product, run the ads, do the outreach, everything.
Struggling with sales now. Thinking of narrowing focus and solving a more specific pain point.
Curious if anyone else here is grinding through early-stage B2B SaaS — would love to hear how you're tackling this.
A few years ago I worked for a company that no longer exists today. They had, among other things, a job search service connected to their ID system. I was also doing my own project at the same time and needed Python developers. I was young and naive and thought I could find a junior and train him quickly. So I posted a vacancy on this job board: "Looking for a Python developer without experience." It turned out that they showed my phone number and it couldn't be turned off.
I received about 3-4 calls from very strange people who demanded to know how to become programmers. For some reason, they all started calling at about 5am. I even gave some useful advice to the first one, because I was taken aback by such impudence.
Today I use about 4 different phone numbers to separate my private life and data leaks like that.