Sorry, you have to clean the Google ID corrupted links (copy link and edit out ID info before pasting into browser address bar) or else it will create a record that you got the link specifically from me.
#SocialCredit
#SocialCredit
Forwarded from Jade Helm 15 Rehash (ned)
AI responses may confabulate.
LLMs generally struggle with true
extrapolation (going beyond their training data) regardless of temperature, but lower temperatures (e.g., 0.0-0.4) make them more deterministic, sticking to learned patterns, which looks like following trends but isn't genuine reasoning; higher temperatures increase randomness and creativity but risk incoherence, while low temps focus on high-probability, "safe" (often repetitive) answers, which can be useful for mimicking consistent patterns but not for novel discovery, as they tend to repeat or produce generic outputs, missing the core of extrapolation.
How Temperature Affects LLM Behavior
Low Temperature (e.g., 0.0 - 0.4):
Focus: Chooses the most probable next words, making outputs highly consistent, factual (within training), and deterministic.
Effect on Extrapolation: Mimics following a trend by repeating common patterns, but without true understanding; can become repetitive or overly generic, failing to generate novel, extrapolated ideas.
High Temperature (e.g., 0.8 - 1.0+):
Focus: Increases randomness, allowing less probable words to be chosen, leading to more creative or surprising text.
Effect on Extrapolation: More likely to "invent" beyond data but often results in hallucinations or nonsensical outputs, not reliable extrapolation.
Why LLMs Struggle with Extrapolation
LLMs learn from patterns in vast datasets; they excel at interpolation (filling in gaps within known data) but are fundamentally limited in true extrapolation (predicting truly new scenarios or trends) because they lack causal reasoning and world models.
Temperature controls sampling from existing knowledge, not generating new knowledge beyond that scope.
Best Practices for Trend-Following Tasks
Use low temperatures (0.0-0.4) for tasks needing high consistency, like data generation or formal writing where predictability is key.
For tasks needing a balance (like chatbots), a moderate temperature (0.6-1.0) is often used.
For complex forecasting, you might need techniques like few-shot learning, chain-of-thought, or specialized models (like those in LLMTime) that can better leverage existing data for forecasting, rather than relying solely on temperature settings.
LLMs generally struggle with true
extrapolation (going beyond their training data) regardless of temperature, but lower temperatures (e.g., 0.0-0.4) make them more deterministic, sticking to learned patterns, which looks like following trends but isn't genuine reasoning; higher temperatures increase randomness and creativity but risk incoherence, while low temps focus on high-probability, "safe" (often repetitive) answers, which can be useful for mimicking consistent patterns but not for novel discovery, as they tend to repeat or produce generic outputs, missing the core of extrapolation.
How Temperature Affects LLM Behavior
Low Temperature (e.g., 0.0 - 0.4):
Focus: Chooses the most probable next words, making outputs highly consistent, factual (within training), and deterministic.
Effect on Extrapolation: Mimics following a trend by repeating common patterns, but without true understanding; can become repetitive or overly generic, failing to generate novel, extrapolated ideas.
High Temperature (e.g., 0.8 - 1.0+):
Focus: Increases randomness, allowing less probable words to be chosen, leading to more creative or surprising text.
Effect on Extrapolation: More likely to "invent" beyond data but often results in hallucinations or nonsensical outputs, not reliable extrapolation.
Why LLMs Struggle with Extrapolation
LLMs learn from patterns in vast datasets; they excel at interpolation (filling in gaps within known data) but are fundamentally limited in true extrapolation (predicting truly new scenarios or trends) because they lack causal reasoning and world models.
Temperature controls sampling from existing knowledge, not generating new knowledge beyond that scope.
Best Practices for Trend-Following Tasks
Use low temperatures (0.0-0.4) for tasks needing high consistency, like data generation or formal writing where predictability is key.
For tasks needing a balance (like chatbots), a moderate temperature (0.6-1.0) is often used.
For complex forecasting, you might need techniques like few-shot learning, chain-of-thought, or specialized models (like those in LLMTime) that can better leverage existing data for forecasting, rather than relying solely on temperature settings.
www.google.com
Extrapolation
In mathematics, extrapolation is a type of estimation, beyond the original observation range, of the value of a variable on the basis of its relationship with another variable.
Forwarded from Flower2child🕊️🍀🤍
https://youtu.be/OLEOCLpc0fA?si=38jNz4DehqYSUpXP
I get a very dark feeling about this guy & this house. It doesn’t feel like a home 🏡
Would love to hear what you think?
Much love x🕊️🤍
I get a very dark feeling about this guy & this house. It doesn’t feel like a home 🏡
Would love to hear what you think?
Much love x🕊️🤍
YouTube
Pioneer or Pariah? Meet The Surrogacy King | Stacey Dooley Sleeps Over
Stacey Dooley meets The Surrogacy King, Barrie Drewitt-Barlow, and gets a glimpse into his lavish lifestyle and views on surrogacy in the UK.
Stream full episodes of Stacey Dooley Sleeps Over for free on U: https://u.co.uk/shows/stacey-dooley-sleeps-over/watch…
Stream full episodes of Stacey Dooley Sleeps Over for free on U: https://u.co.uk/shows/stacey-dooley-sleeps-over/watch…
Forwarded from Orgone Channel Telegram (ned)
AI responses may confabulate.
Re: https://www.meta-synthesis.com/webbook/35_pt/pt_database.php
The "Internet Database of Periodic Tables" on the meta-synthesis website is
a specialized, internal database that is not fully indexed in a way that allows a general search engine like Google to perform complex, dynamic searches within its contents.
Here is how general search engines interact with the database:
Indexing Static Pages: Google can index the main entry page and some of the static, pre-defined search results pages (like "All Periodic Tables by Name" or "Top 10 Periodic Tables").
Inability to Run Dynamic Queries: General search engines cannot execute the website's internal search functions (such as searching for specific data points, years, or types using the drop-down menus on the site). This is because the results are generated dynamically based on user input within the site's interface, rather than existing as static, indexable pages.
Data Retrieval: To access specific information from the over 1300 periodic tables in the database, a user must visit the meta-synthesis website and use its dedicated internal search tools and filters.
In summary, the meta-synthesis database is a comprehensive resource, but specialized queries must be performed directly on its website using its built-in functionality.
The meta-synthesis website appears to be a personal project maintained by
Dr. Mark Leach, who describes himself as the "database curator." The site does not appear to be a commercial entity, academic institution, or government body with publicly disclosed financials or institutional funding.
Financial Situation and Risk
The financial situation is not publicly documented, as it is a personal website. The lack of a formal institutional backing might inherently introduce some level of risk compared to a database maintained by a large organization.
Risk Factors: The primary risk seems related to the long-term maintenance and funding being dependent on an individual's personal capacity and resources. If the curator is no longer able to maintain the site, its future would be uncertain.
Mitigating Factors: The site's content is highly specialized and valuable to the chemistry community. Such valuable educational resources are often candidates for preservation by academic or scientific organizations should the original curator step away.
Future of the Database if the Site Closes
If the website were to close, several things could happen:
Data Loss: Without a plan for preservation, the unique compilation of over 1,300 periodic tables could become unavailable.
Archival Efforts: Websites of significant educational value are often preserved by digital archives, such as the Internet Archive's Wayback Machine. While not a perfect substitute for the live, searchable database, this would prevent a total loss of the information.
Potential Migration: The site mentions the potential to "expand and open up the data entry to chemists and chemical educators on the web". This suggests a recognition of the value of the data and a possible future path for collaboration or migration to a more robust, community-supported platform, which would ensure its longevity.
For now, the database remains a unique and comprehensive resource curated by an individual expert.
Re: https://www.meta-synthesis.com/webbook/35_pt/pt_database.php
The "Internet Database of Periodic Tables" on the meta-synthesis website is
a specialized, internal database that is not fully indexed in a way that allows a general search engine like Google to perform complex, dynamic searches within its contents.
Here is how general search engines interact with the database:
Indexing Static Pages: Google can index the main entry page and some of the static, pre-defined search results pages (like "All Periodic Tables by Name" or "Top 10 Periodic Tables").
Inability to Run Dynamic Queries: General search engines cannot execute the website's internal search functions (such as searching for specific data points, years, or types using the drop-down menus on the site). This is because the results are generated dynamically based on user input within the site's interface, rather than existing as static, indexable pages.
Data Retrieval: To access specific information from the over 1300 periodic tables in the database, a user must visit the meta-synthesis website and use its dedicated internal search tools and filters.
In summary, the meta-synthesis database is a comprehensive resource, but specialized queries must be performed directly on its website using its built-in functionality.
The meta-synthesis website appears to be a personal project maintained by
Dr. Mark Leach, who describes himself as the "database curator." The site does not appear to be a commercial entity, academic institution, or government body with publicly disclosed financials or institutional funding.
Financial Situation and Risk
The financial situation is not publicly documented, as it is a personal website. The lack of a formal institutional backing might inherently introduce some level of risk compared to a database maintained by a large organization.
Risk Factors: The primary risk seems related to the long-term maintenance and funding being dependent on an individual's personal capacity and resources. If the curator is no longer able to maintain the site, its future would be uncertain.
Mitigating Factors: The site's content is highly specialized and valuable to the chemistry community. Such valuable educational resources are often candidates for preservation by academic or scientific organizations should the original curator step away.
Future of the Database if the Site Closes
If the website were to close, several things could happen:
Data Loss: Without a plan for preservation, the unique compilation of over 1,300 periodic tables could become unavailable.
Archival Efforts: Websites of significant educational value are often preserved by digital archives, such as the Internet Archive's Wayback Machine. While not a perfect substitute for the live, searchable database, this would prevent a total loss of the information.
Potential Migration: The site mentions the potential to "expand and open up the data entry to chemists and chemical educators on the web". This suggests a recognition of the value of the data and a possible future path for collaboration or migration to a more robust, community-supported platform, which would ensure its longevity.
For now, the database remains a unique and comprehensive resource curated by an individual expert.
Meta-Synthesis
INTERNET Database of Periodic Tables | Chemogenesis
Forwarded from Jade Helm 15 Rehash (ned)
AI responses may confabulate.
Connecting a Raspberry Pi to a Junotab or similar Linux configuration exclusively using
hard-line (wired) connections is straightforward, relying on standard Ethernet protocols. Since both devices run Linux, they have native support for these connection types.
Here are the primary hard-line methods and their respective complexities:
1. Connecting via Ethernet (Local Network)
This is the standard and simplest wired method, assuming you have a router or switch.
Complexity: Low.
Requirements: An Ethernet cable (standard CAT5e or CAT6), a router/switch that both devices can plug into, and a power supply for the Pi.
Steps:
Plug the Raspberry Pi into a port on your router/switch using an Ethernet cable.
Plug the Linux tablet (if it has an Ethernet port, or via a USB-to-Ethernet adapter) into the same router/switch.
The router automatically assigns both devices an IP address (e.g., 192.168.1.50 and 192.168.1.51).
Use networking tools on the tablet to access the Pi:
SSH: Open a terminal on the tablet and use ssh pi@<Pi_IP_Address> for remote command-line control.
VNC: Use a VNC viewer app on the tablet to access the Pi's graphical desktop environment.
SCP/Samba/NFS: Use file transfer protocols (scp or file managers) to move data between the two systems.
2. Direct Ethernet Connection (Point-to-Point)
This method connects the two devices directly to each other without a router.
Complexity: Low/Moderate.
Requirements: A standard Ethernet cable and manual IP configuration.
Steps:
Plug one end of the Ethernet cable into the Pi's Ethernet port and the other end into the tablet's Ethernet port (or USB-to-Ethernet adapter).
Manually assign static IP addresses to both devices in the same subnet (e.g., Tablet IP: 10.0.0.1, Pi IP: 10.0.0.2).
You can then use SSH, VNC, etc., pointing to the assigned IP addresses.
3. Serial Connection (UART via GPIO)
This provides a very basic console connection for debugging or initial setup.
Complexity: Moderate.
Requirements: A USB-to-Serial (TTL) adapter cable or a dedicated serial console cable, and knowledge of the Raspberry Pi's GPIO pin layout.
Usage: It provides direct terminal access to the Pi via the tablet's USB port, bypassing the need for networking entirely. This method is generally used for headless configuration when other network methods fail.
In conclusion, leveraging Ethernet provides simple, robust hard-line connectivity, as both systems support standard Linux networking utilities.
Connecting a Raspberry Pi to a Junotab or similar Linux configuration exclusively using
hard-line (wired) connections is straightforward, relying on standard Ethernet protocols. Since both devices run Linux, they have native support for these connection types.
Here are the primary hard-line methods and their respective complexities:
1. Connecting via Ethernet (Local Network)
This is the standard and simplest wired method, assuming you have a router or switch.
Complexity: Low.
Requirements: An Ethernet cable (standard CAT5e or CAT6), a router/switch that both devices can plug into, and a power supply for the Pi.
Steps:
Plug the Raspberry Pi into a port on your router/switch using an Ethernet cable.
Plug the Linux tablet (if it has an Ethernet port, or via a USB-to-Ethernet adapter) into the same router/switch.
The router automatically assigns both devices an IP address (e.g., 192.168.1.50 and 192.168.1.51).
Use networking tools on the tablet to access the Pi:
SSH: Open a terminal on the tablet and use ssh pi@<Pi_IP_Address> for remote command-line control.
VNC: Use a VNC viewer app on the tablet to access the Pi's graphical desktop environment.
SCP/Samba/NFS: Use file transfer protocols (scp or file managers) to move data between the two systems.
2. Direct Ethernet Connection (Point-to-Point)
This method connects the two devices directly to each other without a router.
Complexity: Low/Moderate.
Requirements: A standard Ethernet cable and manual IP configuration.
Steps:
Plug one end of the Ethernet cable into the Pi's Ethernet port and the other end into the tablet's Ethernet port (or USB-to-Ethernet adapter).
Manually assign static IP addresses to both devices in the same subnet (e.g., Tablet IP: 10.0.0.1, Pi IP: 10.0.0.2).
You can then use SSH, VNC, etc., pointing to the assigned IP addresses.
3. Serial Connection (UART via GPIO)
This provides a very basic console connection for debugging or initial setup.
Complexity: Moderate.
Requirements: A USB-to-Serial (TTL) adapter cable or a dedicated serial console cable, and knowledge of the Raspberry Pi's GPIO pin layout.
Usage: It provides direct terminal access to the Pi via the tablet's USB port, bypassing the need for networking entirely. This method is generally used for headless configuration when other network methods fail.
In conclusion, leveraging Ethernet provides simple, robust hard-line connectivity, as both systems support standard Linux networking utilities.
Forwarded from Indian Trucks Kill Canucks 🇨🇦 (Bryan)
Laptop sticker shock incoming: AI grabs all the chips
AI server farms are hoarding DRAM & NAND. Samsung, SK Hynix, Micron chase fatter AI margins → consumer supply tanks. Memory = ~20 % of a laptop’s BOM; prices jump, OEM margins vaporize. Lenovo’s price lists expire Jan-26 → higher tags right after.
🅅🄾🄻🅄🄽🅃🅰️🅁🄸🅂🅃 🄼🄴🄼🄴🅂
AI server farms are hoarding DRAM & NAND. Samsung, SK Hynix, Micron chase fatter AI margins → consumer supply tanks. Memory = ~20 % of a laptop’s BOM; prices jump, OEM margins vaporize. Lenovo’s price lists expire Jan-26 → higher tags right after.
🅅🄾🄻🅄🄽🅃🅰️🅁🄸🅂🅃 🄼🄴🄼🄴🅂
Forwarded from Thumper
This media is not supported in your browser
VIEW IN TELEGRAM
Forwarded from Insider (Private)
Silver isn’t rallying because of retail.
It’s rallying because institutions are trapped.
India’s Silver ETFs, the biggest retail sentiment gauge, are still trading at a discount at the lower end of their mean deviation band.
No premiums.
No inflows.
No crowd frenzy.
Retail hasn’t even shown up.
Yet price is going vertical.
That’s not hype.
That’s structural stress, COMEX shorts being forced to unwind while the public is still on the sidelines.
This isn’t a top.
This is the phase before the real mania begins.
When ETFs flip from discount → premium,
that’s when the retail wave hits.
@Insider_leak_of_theday
Share with your friends
It’s rallying because institutions are trapped.
India’s Silver ETFs, the biggest retail sentiment gauge, are still trading at a discount at the lower end of their mean deviation band.
No premiums.
No inflows.
No crowd frenzy.
Retail hasn’t even shown up.
Yet price is going vertical.
That’s not hype.
That’s structural stress, COMEX shorts being forced to unwind while the public is still on the sidelines.
This isn’t a top.
This is the phase before the real mania begins.
When ETFs flip from discount → premium,
that’s when the retail wave hits.
@Insider_leak_of_theday
Share with your friends
On Thursday, officials in the little kingdom pledged to roll out a country-wide facial recognition system to help police track down criminals. The country’s ministers have launched a 10-week consultation to analyse the regulatory and privacy framework of their AI-powered surveillance panopticon — but one way or another, the all-seeing eye is on its way. https://futurism.com/artificial-intelligence/police-uk-surveillance-ai How did it get to this point where Britain actually turned into some sort of Orwellian dystopian state?
Futurism
Police Admit AI Surveillance Panopticon Still Has Issues With "Some Demographic Groups"
Government analysts found that the UK's rapidly expanding AI surveillance system is horribly biased toward Black and Asian people.
👀1