Dylan Layne Tanner
@DylanLTanner
fixing local problems with digital tools | applied ai | founder @amethron_
Вам может понравиться
Flock called us "terrorists." Palantir called us "parasites." Surveillance bros fear surveillance....Ironic. If transparency threatens your entire business model, that should tell us everything we need to know. 🚩
A number of people are talking about implications of AI to schools. I spoke about some of my thoughts to a school board earlier, some highlights: 1. You will never be able to detect the use of AI in homework. Full stop. All "detectors" of AI imo don't really work, can be…
Gemini Nano Banana Pro can solve exam questions *in* the exam page image. With doodles, diagrams, all that. ChatGPT thinks these solutions are all correct except Se_2P_2 should be "diselenium diphosphide" and a spelling mistake (should be "thiocyanic acid" not "thoicyanic") :O
We have to, though, if we want to "win" over the long term. That's the paradox. We're in this mess because of circular financing, AGI hype, and a focus on big, closed-source resource hogs. More overinvestment just leads to a harder landing down the road. Best way out would be…
If AI (or any form of automation) can reduce or eliminate the need for certain types of work, that’s good. But under the system we have right now, the gains from increased productivity will only accrue to finance capital. The rest of us will be treated as expendable. How this…
when ppl say “ai will take jobs,” how do you not read that as a gigantic W? like i desperately want ai to take my job. the whole arc of tech & civilization in general gas been to delete labor so you only work when you want to, not because you need to grind for survival.…
This is such a common sentiment among developers (& one I share). So much of the smart device / IoT thing has been a solution in search of a problem. We've taken a lot of simple, fault-tolerant devices and introduced failure points and levels of complexity that didn't need to be…
As a software developer, I want software out of almost everything I use
I've used Linux for ages as a programmer (remote, via Vagrant, or via WSL), but a couple of days ago I said "screw it" and decided to switch to it for everything else, too. Man...it's come a long ways. @fedora Workstation has been a massive upgrade. Easy to get set up + slick…
Microsoft admits File Explorer is slow in Windows 11, and it’s going to preload it in the background to help improve launch performance. “This shouldn’t be visible to you, outside of File Explorer hopefully launching faster when you need to use it,” Microsoft confirmed. If you…
AI illiteracy + general illiteracy + lack of critical thinking: unfortunately, the perfectly manipulable society
This is a really important point. LLMs are reductive. Telling you "what is known to be" or "what is likely to be" isn't a substitute for the way human researchers establish a hypothesis and apply actual reasoning. This isn't the same as saying "LLMs are useless for serious…
A lot of people think GPT-5 generalizes out of distribution (OOD) because it can help, in some very limited form, in scientific research. Yet usually (99.99% of the time) what GPT-5 is normally doing is searching and surfacing existing literature not solving novel problems.
Once I understand how the magic trick is done (transformers + tool use), three things happened: - I started to like programming & machine learning again (after getting burned out) - I gained a immense respect for how efficiently and mysteriously the human mind works - I…
The Windows AI / Copilot pushback is interesting because it highlights the difference between optimizing for interaction and optimizing for orchestration. I don't need to be able to talk to my operating system. With a good UI, I can do what I need faster than cloud-based AI can.…
The notion that AGI would have infinite returns has been used to justify investment far above expected returns (by 10x-100x) for technology that is neither AGI nor on the path to AGI
So the problem isn't AI—just as it wasn't industrial robotics or electrification or Newcomen's steam engine... The problem is the mode of production itself: an economic system where huge swaths of the population engage in unproductive labor, virtually nothing is spent on…
📁 Geoffrey Hinton, a pioneer of AI, warns that workers who lose their jobs may have nowhere to go because AI will perform any task humans can. He notes that roles like call center agents will vanish since AI will do their work better, faster and cheaper. He says the employment…
Most sites behind @Cloudflare experience *way more downtime* from CF's own failures than they ever would from a hypothetical DDoS attack Your tiny blog isn't getting nation-state DDoSed, but it *is* getting Cloudflared every few months! 😂
Somewhere along the way a lot of us devs got sold on the idea that Cloudflare is just "part of the stack." Pages for static sites / frontends, Workers for services, the CDN for just about everything But a lot of times, it's overkill. We launched a couple of projects without…
There are a bunch of things going on here under the surface: - They're banging the "AI safety" drum (again) - They're selling to the cybersecurity crowd - They're quietly showcasing Claude's capabilities as an attack tool to government agencies with fat wallets "We also advise…
We disrupted a highly sophisticated AI-led espionage campaign. The attack targeted large tech companies, financial institutions, chemical manufacturing companies, and government agencies. We assess with high confidence that the threat actor was a Chinese state-sponsored group.
Like I said, Flashforge got a threat letter and responded poorly. Not the best PR response, but it'll have to do. When I've checked in OPNsense, the printers (mainly Flashforge and Anycubic) are some of the least suspicious things connected to the network.
I’ll take an unpopular stance on this: Flashforge is probably doing the bare minimum to comply with the Manhattan DA’s threat letter. They’re an overseas printer manufacturer. It isn’t worth it to them to take a libertarian stand and lose access to a big chunk of their market.…
Why I'm betting on China in the AI race. If you aren't aware, the only differentiator between the trillion-parameter size frontier models is the dataset on which they have been finetuned. Claude has been the best in debugging not because Amodei invented some original way of…
United States Тренды
- 1. National Guard 313K posts
- 2. Thanksgiving 497K posts
- 3. Afghan 42.7K posts
- 4. Cease 21.7K posts
- 5. Blue Jays 8,479 posts
- 6. Blood 194K posts
- 7. Arsenal 507K posts
- 8. Liverpool 162K posts
- 9. Rahmanullah Lakanwal 4,833 posts
- 10. Derrick White N/A
- 11. #StrangerThings5 28.5K posts
- 12. Slot 132K posts
- 13. Al Sharpton 5,000 posts
- 14. Frank Ragnow 9,119 posts
- 15. Bayern 260K posts
- 16. Seditious Six 141K posts
- 17. #triplegobble N/A
- 18. Neuer 26.9K posts
- 19. Tony Brothers N/A
- 20. Insurrection Act 9,527 posts
Вам может понравиться
Something went wrong.
Something went wrong.