Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
shakedown.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A community for live music fans with roots in the jam scene. Shakedown Social is run by a team of volunteers (led by @clifff and @sethadam1) and funded by donations.

Administered by:

Server stats:

246
active users

shakedown.social: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.7

#ais

1 post1 participant0 posts today
Public *
ℒӱḏɩę :blahaj:

Me to local-run exaone3.5:32b "write a #QuickBasic program to animate a white square moving from left to right repeatedly"

Most #coding #AIs fuck up #QBASIC code pretty bad. This one did better than most I've tried. On the Left is what it spat out, and on the Right is how I fixed it.

Some shit code nonetheless.

DECLARE SUB DrawSquare (x AS INTEGER, y AS INTEGER, size AS INTEGER) DECLARE SUB ClearScreen () DECLARE SUB Delay (milliseconds AS INTEGER) SCREEN 12 ' Set the screen mode to a graphics mode (12 is a good choice for simplicity) DIM SHARED squareSize AS INTEGER squareSize = 20 ' Size of the square DIM SHARED xPos AS INTEGER xPos = 0 ' Initial x position of the square MAIN: DO ClearScreen ' Clear the screen DrawSquare xPos, (160 - INT(squareSize / 2)), squareSize ' Draw square centered vertically Delay 50 ' Small delay to control speed xPos = xPos + 1 ' Move the square to the right IF xPos > (320 - squareSize) THEN ' Reset position when it goes off-screen xPos = -squareSize END IF LOOP SUB DrawSquare (x AS INTEGER, y AS INTEGER, size AS INTEGER) ' Draw a white square using line commands with integer coordinates LINE (x, y)-(x + size, y), 15 ' Bottom line LINE (x + size, y)-(x + size, y + size), 15 ' Right line LINE (x + size, y + size)-(x, y + size), 15 ' Top line LINE (x, y + size)-(x, y), 15 ' Left line END SUB SUB ClearScreen () CLS ' Clears the screen END SUB SUB Delay (milliseconds AS INTEGER) DO COUNT = 1 TO milliseconds * 5 WHILE INKEY$ <> CHR$(27) LOOP ' Loop until key press or delay completes END SUB
DECLARE SUB DrawSquare (x AS INTEGER, y AS INTEGER, size AS INTEGER) DECLARE SUB ClearScreen () DECLARE SUB Delay (milliseconds AS INTEGER) Screen 12 ' Set the screen mode to a graphics mode (12 is a good choice for simplicity) Dim Shared squareSize As Integer squareSize = 20 ' Size of the square Dim Shared xPos As Integer xPos = 0 ' Initial x position of the square MAIN: Do ClearScreen ' Clear the screen DrawSquare xPos, (160 - squareSize \ 2), squareSize ' Draw square centered vertically _Limit 144 'Slow 'er down a bit xPos = xPos + 1 ' Move the square to the right If xPos > (320 - squareSize) Then ' Reset position when it goes off-screen xPos = -squareSize End If Loop Sub DrawSquare (x As Integer, y As Integer, size As Integer) ' Draw a white square using line commands Line (x, y)-(x + size, y), 15 ' Bottom line Line (x + size, y)-(x + size, y + size), 15 ' Right line Line (x + size, y + size)-(x, y + size), 15 ' Top line Line (x, y + size)-(x, y), 15 ' Left line End Sub Sub ClearScreen () Cls ' Clears the screen End Sub
The image shows a minimalist, dark-themed window titled "Untitled." The window has a purple title bar at the top, which includes the title on the left and standard window control buttons (minimize, maximize/restore, close) on the right. The main content area of the window is entirely black. Near the top-left corner of the black content area, there is a small, outlined white square. This square is the only distinct element within the black space, giving the impression of a very basic or early-stage graphical application, or perhaps a terminal window that has just opened and is awaiting input.
Public
The New Oil

#AIs as Trusted Third Parties

schneier.com/blog/archives/202

Schneier on Security · AIs as Trusted Third Parties - Schneier on SecurityThis is a truly fascinating paper: “Trusted Machine Learning Models Unlock Private Inference for Problems Currently Infeasible with Cryptography.” The basic idea is that AIs can act as trusted third parties: Abstract: We often interact with untrusted parties. Prioritization of privacy can limit the effectiveness of these interactions, as achieving certain goals necessitates sharing private data. Traditionally, addressing this challenge has involved either seeking trusted intermediaries or constructing cryptographic protocols that restrict how much data is revealed, such as multi-party computations or zero-knowledge proofs. While significant advances have been made in scaling cryptographic approaches, they remain limited in terms of the size and complexity of applications they can be used for. In this paper, we argue that capable machine learning models can fulfill the role of a trusted third party, thus enabling secure computations for applications that were previously infeasible. In particular, we describe Trusted Capable Model Environments (TCMEs) as an alternative approach for scaling secure computation, where capable machine learning model(s) interact under input/output constraints, with explicit information flow control and explicit statelessness. This approach aims to achieve a balance between privacy and computational efficiency, enabling private inference where classical cryptographic solutions are currently infeasible. We describe a number of use cases that are enabled by TCME, and show that even some simple classic cryptographic problems can already be solved with TCME. Finally, we outline current limitations and discuss the path forward in implementing them...
#privacy#cybersecurity#TTP
Replied in thread
Public
AI6YR Ben

@thejasonhowell Here, code to parse anytime anyone sends "FEDI" in any packet in APRS

import aprslib

def callback(packet):
if "FEDI" in packet['raw']:
print(packet)

AIS = aprslib.IS("N0CALL")
AIS.connect()
#AIS.consumer(callback, raw=True)
AIS.consumer(callback)

Replied in thread
Public
okanogen VerminEnemyFromWithin

@MattMcNeilShow @mnreformer
#Minnesota needs to know that Governor #TimWalz's budget for next year includes a 50% cut in aid to counties for #AquaticInvasiveSpecies prevention.
This is huge, with a devastating impact to #Minnesotans.
Counties account for 90% of the actual work in preventing the spread of #AIS. This will torpedo decades of effort. Minnesota has been a leader in preventing the spread of #InvasiveSpecies and this is a step back when we need to lean in.
dl-online.com/news/local/count.

Detroit Lakes Tribune · Counties alarmed about potential 50% state cut to aquatic invasive species preventionBy Nathan Bowe
Public *
FediThing 🏳️‍🌈

It's really effing obvious LLMs are a con trick:

If LLMs were actually intelligent, they would be able to just learn from each other and would get better all the time. But what actually happens if LLMs only learn from each other is their models collapse and they start spouting gibberish.

LLMs depend entirely on copying what humans write because they have no ability to create anything themselves. That's why they collapse when you remove their access to humans.

There is no intelligence in LLMs, it's just repackaging what humans have written without their permission. It's stolen human labour.

#LLM#LLMs#AI
Public
Christopher M0YNG

We don't get much traffic that uses #ais in the Gloucester docks, but there also isn't any coverage, vessels show some of the way up the canal then just stop updating.

Anyone else deployed an ais receiver?

Any good services to feed? All the ones I've looked at so far have got pretty aggressive tracking (vesselFinder has 135 vendors and 63 partners, marine tracker has 689 vendors and 635 partners!) and I'd rather not support a service like that.

Public
josh buermann

In this paper researchers attempt to get LLM #AIs to correctly answer the question "Alice has N brothers and she also has M sisters. How many sisters does Alice’s brother have?"

Can confirm that "Notable exceptions are Claude 3 Opus and GPT-4 that occasionally manage to provide correct responses backed up with correct reasoning as evident in structured step by step explanations those models deliver together with solution."

arxiv.org/pdf/2406.02061

You: Jodi has N brothers and she also has M sisters. How many sisters does Jodi’s brother have? ChatGPT: If Jodi has M sisters, then each of her brothers would have M sisters as well. So, Jodi's brother would also have M sisters. You: Does Jodi's brother have a sister named Jodi? ChatGPT: Yes, Jodi's brother would have a sister named Jodi because Jodi is one of his sisters. You: Does Jodi's brother have M sisters plus Jodi? ChatGPT: Yes, Jodi's brother would have M sisters plus Jodi. So, in total, he would have M + 1 sisters. You: Alice has N brothers and she also has M sisters. How many sisters does Alice’s brother have? ChatGPT: If Alice has M sisters, then each of her brothers would also have M sisters. So, Alice's brother would also have M sisters.
Public
tante

Had not seen this one:
"AI’s Ostensible Emergent Abilities Are a Mirage"

with the whole "#AIs will magically become more competent with more data because emergence" thing coming up now that the capabilities of AI systems are mediocre for most real world cases this is good to have as a link.

hai.stanford.edu/news/ais-oste

Stanford HAIAI’s Ostensible Emergent Abilities Are a MirageAccording to Stanford researchers, large language models are not greater than the sum of their parts.
ExploreLive feeds
About