iiinova — where your digital self keeps evolving

rphle @alpharee.mastodon.social.ap.brid.gy · 20/06/2026

Wouldn't have happened in the EU

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 01/04/2026

I'm glad they put that orange cone in front of it, otherwise people wouldn't notice the rocket.

Close-up view of the base of a large rocket on a launch pad, showing multiple engine nozzles beneath an orange central tank and two white side boosters, with an orange safety cone visible in the foreground.

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 12/10/2025

Today I learned 'bool' is a subclass of 'int' in Python. Before Python introduced a boolean type, it used 0 and 1 for truth values. 'bool' was later designed to behave like 0 and 1 to maintain backward compatibility. #python #programming #python2 #python3 #til

>>> isinstance(True, int)
True
>>> issubclass(bool, int)
True

101 1

rphle @alpharee.mastodon.social.ap.brid.gy · 07/10/2025

The shape of the Earth remains an active area of scientific research.

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 16/07/2025

We're seriously wasting energy so AI can place random guesses. #google #aisummary #slop #ai #rant #duckduckgo

Google AI Summary for "is duckduckgo down":
"No, DuckDuckGo is not down. User reports indicate no current problems with the search engine. IsDown.app and Downdetector.co.uk both show that DuckDuckGo is currently operational."

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 26/05/2025

"I'm only a language model and don't have the capacity to understand and respond." 😩

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 18/05/2025

I love playing around with Google Trends. 😄 Here’s Google search interest for “remove president” from 2020 to 2025. #googletrends #google #removepresident #president #trump #biden #datascience #dataisbeautiful #visualization #politics #jan6 #graph #plot #matplotlib

A matplotlib plot titled "Google Search Interest Over Time: 'Remove President' (US)" shows the relative search interest for the phrase "Remove President" in the United States from July 2020 to May 2025. The y-axis represents search interest, while the x-axis spans from July 2020 to May 2025. The graph features a colored area under the curve: red shading highlights periods around significant political events, while blue shading covers the rest.

Key political events are marked with vertical dashed lines and labels, including:
- 2020 Election (Nov 2020)
- Capitol Riot (Jan 2021)
- Biden Inauguration (Jan 2021)
- Afghanistan Withdrawal (Aug 2021)
- Jan 6 Committee Report (Dec 2022)
- Trump Assassination Attempt (Jul 2024)
- 2024 Election (Nov 2024)
- Trump 2nd Inauguration (Jan 2025)
- "Liberation Day" Tariffs (Mar 2025)

Notable spikes in search interest occur around the 2020 Election, Capitol Riot, Biden Inauguration, and again after the 2024 Election and Trump’s 2nd Inauguration. The graph visually links peaks in search activity to these political milestones, suggesting increased public concern or controversy during these times.

014 4

rphle @alpharee.mastodon.social.ap.brid.gy · 26/04/2025

I don't usually post for #Caturday because I don't have a cat myself — but luckily, other people do! Spotted in the window of an antiquarian bookshop in Wrocław, Poland.

A white cat sitting in the window of an antiquarian bookshop on top of some books. In the window reflection, you can see some buildings.

040 4

rphle @alpharee.mastodon.social.ap.brid.gy · 25/04/2025

Make Python imports great again. 🔥

$PIP package page: 👊 TARIFF 🔥 The GREATEST, most TREMENDOUS Python package that makes importing great again! About TARIFF is a fantastic tool that lets you impose import tariffs on Python packages. We're going to bring manufacturing BACK to your codebase by making foreign imports more EXPENSIVE! Installation pip install tariff Usage import tariff # Set your tariff rates (package_name: percentage) tariff.set({ "numpy": 50, # 50% tariff on numpy "pandas": 200, # 200% tariff on pandas "requests": 150 # 150% tariff on requests }) # Now when you import these packages, they'll be TARIFFED! import numpy # This will be 50% slower import pandas # This will be 200% slower$

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 02/04/2025

I started building a small, mostly LLM driven Mastodon bot today. I'm pretty happy with how it's performing, here's a little preview, I'm aiming to publish it by Sunday. Can you guess what it's about? #llm #mastodonbot #preview #bot #ai #cool #headline #news

Headline: "Wikipedia Partners with ICE Air to Deport AI Bots via 'Secure' Gmail Channels"
Abstract: "Wikipedia has teamed up with ICE Air to deport rogue AI scrapers while maintaining its commitment to transparency, routing all deportation orders through Gmail to ensure every 'confidential' memo becomes public domain..."

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 23/03/2025

Wow! Just tested EXAONE Deep (2.4B, 7.8B, 32B) and I'm impressed. It's really fast and great at reasoning: Even the 2.4B model solved "Count to ten, skipping numbers ending in 'e'" right away, which all ≤32B models I tested before failed. It does seem to […] [Original post on mastodon.social]

>>> Count from 0 to ten while skipping every number ending in 'e'.

<thought>
[...]
</thought>
To solve the problem of counting from 0 to ten while skipping numbers whose spelled-out form ends with the letter 'e',
we first list each number from 0 to 10 and check the last letter of their English spellings:

- **0 (zero)**: Ends with 'O' → keep
- **1 (one)**: Ends with 'E' → skip
- **2 (two)**: Ends with 'O' → keep
- **3 (three)**: Ends with 'E' → skip
- **4 (four)**: Ends with 'R' → keep
- **5 (five)**: Ends with 'E' → skip
- **6 (six)**: Ends with 'X' → keep
- **7 (seven)**: Ends with 'M' → keep
- **8 (eight)**: Ends with 'T' → keep
- **9 (nine)**: Ends with 'E' → skip
- **10 (ten)**: Ends with 'N' → keep

The numbers retained are: 0, 2, 4, 6, 7, 8, and 10.

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 01/02/2025

People are proclaiming #GlobalSwitchDay and I'm a bit late to the party but this just came to my mind! Drop all of these expensive and data hungry LLM APIs and run Large Language Models on your own computer! ✊ #ai #llm #switch #ollama #openai #anthropic #chatgpt #privacy #local

#GlobalSwitchDay

LLM APIs (Perplexity, Copilot, Mistral, Gemini, Claude, ChatGPT) => Ollama

004 4

rphle @alpharee.mastodon.social.ap.brid.gy · 20/01/2025

Digital archaeology in action: ELIZA, the world's first chatbot from 1964, has been resurrected by MIT researchers after 60 years! The team created an emulator to run ELIZA on a simulated IBM 7094, preserving 96% of the original 2,600 lines of code. They […] [Original post on mastodon.social]

RUN Sample Run HI! I'M ELIZA« WHAT'S YOUR PROBLEM?

? I AM AN IBM 370

DID YOU COME TO ME BECAUSE YOU ARE AN IBM 370

? YES I WANT A NEW LINEPRINTER

WHAT WOULD IT MEAN TO YOU IF YOU GOT A NEV LINEPRINTER ? ABOUT 1200 LPM

SAY» DO YOU HAVE ANY PSYCHOLOGICAL PROBLEMS?

? WHAT WOULD YOU KNOW I°'M A 370 AND YOU'RE AN IMSAI WHAT MAKES YOU THINK I AM AN IMSAL

? IT SAYS THAT ON YOUR FRONT PANEL

WHY ARE YOU CONCERNED ABOUT MY FRONT PANEL

? LOOK YOU BROUGHT UP THE SUBJECT

WE WERE DISCUSSING YOU=- NOT ME.

? 1'M FIXING A HOLE««s

DID YOU COME TO ME BECAUSE YOU ARE FIXING A HOLE«s. ? SORRY THAT WAS A TEN YEAR OLD SOFTWARE BUG PLEASE DON'T APOLOGIZE!

7 1 AM ALWAYS APOLOGIZING

HOW LONG HAVE YOU BEEN ALWAYS APOLOGIZING

7 SINCE MY FIRST COLD START

WHAT DOES THAT SUGGEST 70 YOU?

? YOU'RE THE SHRINK

DOES IT PLEASE YOU TO BELIEVE I AM THE SHRINK

? DONT YOU KNOW?

OH» 1 KNOW?

7 1 AM BEGINNING TO WONDER ABOUT YOU

DO YOU BELIEVE IT IS NORMAL TO BE BEGINNING TO WONDER ABOUT I ? HA HA HA YOU CAN'T EVEN TALK PROPERLY

YOU'RE NOT REALLY TALKING ABOUT ME, ARE YOU?

7 1S THERE ANYONE ELSE IN THE ROOM?

1 SEE.

7 COMPUTERS DONT HAVE EYES

ARE YOU SURE?

7 ABOUT COMPUTERS?

DO COMPUTERS WORRY YOU?

7 SHOULD THEY AFTER ALL I AM A 370

DO YOU ENJOY BEING A 370

?

10 REM

20 REM EL1ZA/DOCTOR

30 REM CREATED BY JOSEPH WEIZENBAUM istil

0 REM THIS VERSION BY JEFF SHRAGER Program Listing

50 REM EDITIED AND MODIFIED FOR MITS 8K BASIC 4.0 BY STEVE NORTH

60 REM CREATIVE COMPUTING PO BOX 789-M MORRISTOWN NJ 07960

70 REM

8 REM oeeee INITIALIZATION=====

90 DIM C3C72),18¢72),KS$C72),F$CT2), SSCT2),RSC72),PSC72),28¢72)

100 DIM 5(36)»R(36),N(36)

110 N1=361N2=121N3=112

120 FOR X=1 TO N1+N2+N3ItREAD ZSINEXT XtREM SAME AS RESTORE

130 FORX=1 TO N1

140 READ S(X),LIR(X)=SCX)INCXI=S(XI4L-1 -

150 NEXT X

160 PRINT "HI1 1'M ELIZA. WHAT'S YOUR PROBLEM?"

170 REM

180 REM ====~USER INPUT SECTION-==-=

190 REM

200 INPUT 1s

201 Ise® "aIsen »

210 REM GET RID OF APOSTROPHES

220 FOR L=l TO LENCIS)

230 IFMIDS(ISsL,1)" '"THENISaLEFTSCIS,L-1)+RIGHTSC 15, LENCI$)=L)1G0T0230

240 IFL+4<=LENCIS)THENIFMIDS(IS,L,4)=""SHUT" THENPRINT"'SHUT UPss+"$END 100 CREATIVE COMPUTING

010 1

rphle @alpharee.mastodon.social.ap.brid.gy · 14/01/2025

Since the trained model has been open-sourced, you can easily run it on your local machine. An unofficial version is also available on ollama: 'ollama run medragondot/Sky-T1-32B-Preview'. More details can also be found on HuggingFace: huggingface.co/NovaSky-AI/Sky-T1-32…

>>> Please count to ten, skipping any number ending in 'e'.
Sure, here's how that would go:

1. One
2. Two
3. Four
4. Six
5. Seven
6. Eight

Wait a minute, I need to clarify the rule. The numbers to skip are those that end with the letter 'e' when spelled out in
English. Let's list the numbers from one to ten and see which ones end with 'e':

- One (ends with e)
- Two (ends with o)
- Three (ends with e)
- Four (ends with r)
- Five (ends with e)
- Six (ends with x)
- Seven (ends with n)
- Eight (ends with t)
- Nine (ends with e)
- Ten (ends with n)

So, the numbers to skip are: one, three, five, nine.

Therefore, counting to ten while skipping those numbers:

1. Two
2. Four
3. Six
4. Seven
5. Eight
6. Ten

So, the sequence is: two, four, six, seven, eight, ten.

000 0

rphle @alpharee.mastodon.social.ap.brid.gy · 14/01/2025

A team at UC Berkeley has trained Sky-T1-32B-Preview, a reasoning model that matches the performance of O1-Preview on popular reasoning and coding benchmarks. What makes this even more remarkable is that the training was completed in 19 hours for less than […] [Original post on mastodon.social]

Sky-T1-32B-Preview Qwen-2.5-32B-Instruct QwQ o1-preview
Math500 82.4 76.2 85.4 81.4
AIME2024 43.3 16.7 50.0 40.0
LiveCodeBench-Easy 86.3 84.6 90.7 92.9
LiveCodeBench-Medium 56.8 40.8 56.3 54.9
LiveCodeBench-Hard 17.9 9.8 17.1 16.3
GPQA-Diamond 56.8 45.5 52.5 75.2

106 6