Key concepts in gamification:

• Reward systems
• Progress tracking
• Narrative and storytelling
• Social engagement
• Game psychology

According to Self-Determination Theory by Csíkszentmihályi, there are three basic psychological needs that drive human behavior (and intrinsic motivation):

• Autonomy
• Competence
• Relatedness

The top 10 core competencies of product managers:

• Qualitative market research
• Quantitative market research
• Facilitating design-thinking workshops
• Prototyping and validation
• Feature prioritization and roadmap planning
• Resource allocation
• Strategic thinking and problem solving
• Translator between the business and technology teams
• Pricing strategy and revenue modeling
• Defining and tracking success metrics

The six fundamental ways in which people are influenced:

• Reciprocity
• Commitment and consistency
• Social proof
• Authority
• Liking
• Scarcity

Examples of decision-making biases:

• Loss aversion
• Social proof
• Status quo bias

Gamification frameworks and methodologies:

• MDA (Mechanics, Dynamics, Aesthetics)
• Fogg Behavior Model (Motivation, Ability, Trigger)
• SCARF Model (Status, Certainty, Autonomy, Relatedness, Fairness)
• HEXAD User Types (Achiever, Socializer, Humanist, Explorer, Architect, Director, Philanthropist, Gambler, Collector, Role-player, Manipulator, Thrill-seeker)
• Octalysis Framework (Epic Meaning and Calling, Development and Accomplishment, Empowerment of Creativity and Feedback, Ownership and Possession, Social Influence and Relatedness, Scarcity and Impatience, Unpredictability and Curiosity, Loss and Avoidance)
• Hooked Model (Trigger, Action, Variable Reward, Investment)
• RAMP framework (Rewards, Achievements, Missions, Progress)
• SAPS Reward Framework (Status, Access, Power, Stuff)
• RECIPE Framework (Reflection, Engagement, Choice, Information, Play, Exposition)
• Game Thinking Framework (Player Insight, Game Design, Smart Feedback, Rapid Iteration)
• PBL Model (Player, Behavior, Loop)
• 6D Framework (Define business objectives, Delineate target behaviors, Describe your players, Devise activity cycles, Don’t forget the fun, Deploy the appropriate tools for the job)
• 4 Keys 2 Fun (Hard Fun, Easy Fun, Serious Fun, People Fun)

Game mechanics - use voting. Each user may add +1 to idea, content or even other users. More advanced (levelled up) users could have more voting power.

Design psychology principles:

• Gestalt Design Psychology
• Dieter Rams’ Ten Principles for Good Design
• Jakob Nielsen’s Heuristics for Interaction Design
• Don Norman’s Principles of Interaction Design

Understanding your user - JTBD - Jobs To Be Done. PM idea: structure Projec Requirements Document (PRD) with JTBD.

Ideation process: Six Thinking Hats (Loyal, Optimism, Critic, Emotion, Creation, Management)

RICE scoring for feature prioritization:

Reach x Impact x Confidence / Effort = RICE Score

Badges naming ideas:

• Ace
• Advocate
• Analyst
• Architect
• Author
• Boss
• Champion
• Commander
• Connoisseur
• Craftsman
• Creator
• Decider
• Detective
• Expert
• Genius
• Guru
• Hero
• Leader
• Legend
• Magician
• Master
• Motivator
• Ninja
• Optimizer
• Pioneer
• Pro
• Resolver
• Rockstar
• Savant
• Specialist
• Strategist
• Tracker
• Virtuoso
• Whiz
• Whisperer
• Wizard

Idea from Strava: You can win medals in certain segments, and when I know I am close to beating a time in a segment, it motivates me to push myself harder.

Idea from Duolingo: Chests may contain rewards such as double XP for the next 15 minutes, a reward that encourages users to keep learning for another 15 minutes, earning extra XP to climb the leaderboard, but also keeping the user learning for 15 more minutes.

Technology drivers of change:

• Quantum computing
• AI and machine learning
• Biotechnology and gene editing
• Augmented and virtual reality (AR/VR)
• Advanced robotics and automation

More to read or watch:

• Boskamp, Elsie, 25 Gamification Statistics 2023: Facts + Trends You Need to Know, Zippia,2022
• Hunicke, Robin; LeBlanc, Marc; and Zubek, Robert, MDA: A Formal Approach to Game Design and Game Research, 2001
• Fogg, BJ, Persuasive Technology: Using Computers to Change What We Think and Do, Morgan Kaufmann, 2003
• Rock, David, SCARF: A Brain-Based Model for Collaborating with and Influencing Others, 2008
• Chou, Yu-Kai, Actionable Gamification: Beyond Points, Badges and Leaderboards, 2015
• Eyal, Nir, Hooked: How to Build Habit-Forming Products, Portfolio, 2014
• Marczewski, Andrzej, The Gamification Design Handbook: Even Ninja Monkeys Like to Play, 2023
• Zichermann, Gabe; Cunningham, Christopher, Gamification by Design: Implementing Game Mechanics in Web and Mobile Apps, O’Reilly Media, 2011
• Werbach, Kevin; Hunter, Dan, For the Win, Revised and Updated Edition: The Power of Gamification and Game Thinking in Business, Education, Government, and Social Impact, Wharton School Press, 2020
• Kim, Amy Jo, Game Thinking: Innovate smarter & drive deep engagement with design techniques from hit games, 2018
• Deterding, Sebastian; Khaled, Rilla; Nacke, Lennart; and Dixon, Dan. Gamification: Toward a Definition, 2011
• Lazzaro, Nicole, Why We Play Games: Four Keys to More Emotion Without Story, 2004
• Nicholson, Scott, A RECIPE for Meaningful Gamification, 2015
• McGonigal, Jane, Reality Is Broken: Why Games Make Us Better and How They Can Change the World, Penguin Books, 2011
• Hodent, Celia, The Gamer’s Brain: How Neuroscience and UX Can Impact Video Game Design, 2017
• Csikszentmihalyi, Mihaly, Flow: The Psychology of Optimal Experience, Harper Perennial Modern Classics
• Super Metroid on TEDx
• Schwarz, M., & Blessing, L., The Design Thinking Toolbox: A Guide to Mastering the Most Popular and Valuable Innovation Methods, Wiley, 2020.
• Gray, D., Brown, S., & Macanufo, J, Gamestorming: A Playbook for Innovators, Rulebreakers, and Changemakers, O’Reilly Media, 2010.
• McGonigal, J, Reality Is Broken: Why Games Make Us Better and How They Can Change the World, Penguin Books, 2011.

“The Famous Computer Cafe” is a radio show from 1980s about the new thing - home computers. 53 episodes have been recovered and are now archived on archive.org. Included are interviews with people like Douglas Adams (The Hitchhiker’s Guide to the Galaxy), Bill Gates, Jack Tramiel (Atari)

Full list of interviews:

List of a developer goals from IEEE article: Measuring Developer Goals.

Information gathering:

• Ensure documentation is up to date
• Understand the context to complete a work item
• Explore the technical solutions (e.g. bugs, design)
• Finding information (e.g. documentation, codelabs, API examples)
• Find an expert

Plan and track work, and manage approvals:

• Know what to work on next
• Coordinate work with peers
• Ensure my launch complies with legal, privacy, and security requirements
• Have my cross-functional team aligned on launch readiness
• Get my design approved
• Design and document a considered plan

Develop, test and commit code:

• Write high quality code
• Ensure the code contributed by others (e.g. teammates, AI) is high quality
• Understand the behaviour of existing code
• Create or maintain holistic test coverage
• Investigate unexpected behavior locally
• Integrate new tools / technology into existing services and systems

Experiment, release and rollout:

• Safely roll out changes to production (e.g. features, models, new releases)
• Run an experiment
• Analyze experiment results

Monitoring, reliability and configuring infrastructure:

• Ensure my product stays within SLO commitments
• Investigate issues in production (e.g. crashes, outages, unexpected behavior)
• Improve system performance
• Ensure my builds stay greem (e.g. build gardening rotations)
• Improve reliability and avoid production problems

Data management:

• Ensure data I’m responsible for is fresh, reliable and of high quality
• Develop and manage data processing pipelines
• Ensure data I’m responsible for is secure and complies with regulations
• Analyze, visualize and understand data to generate insights

Install dependencies:

sudo apt update
sudo apt-get install -y git python3.10 libboost-all-dev jackd2 libjack-jackd2-dev libasound2-dev libpulse-dev libglibmm-2.4-dev libsndfile1-dev libcurl4-gnutls-dev libarchive-dev liblo-dev libtaglib-cil-dev libtaglib-ocaml-dev vamp-plugin-sdk librubberband-dev libudev-dev libfftw3-dev libaubio-dev libxml2-dev libcppunit-dev libusb-1.0-0-dev libwebsockets-dev libcwiid-dev libpangomm-1.4-dev liblrdf0-dev libsamplerate0-dev lv2-dev libserd-dev libsord-dev libsratom-dev liblilv-dev libsuil-dev libsoundtouch-dev libgtkmm-2.4-dev

Clone the source code and build:

git clone https://github.com/Ardour/ardour.git
cd ardour
./waf configure --optimize --lv2 --with-backends=jack,pulseaudio,dummy,alsa
# build using 4 cores, adjust for your system:
./waf build -j4

To run ardours, you don’t need to install it anywhere, simply run in from the same directory:

./gtk2_ardour/ardev

Enjoy!

Enable src packages - edit /etc/apt/sources.list and un-comment lines starting with dev-src.

Install dependencies:

sudo apt update
sudo apt build-dep wesnoth
sudo apt install libcurl4-openssl-dev libsdl2-dev libsdl2-image-dev libsdl2-mixer-dev libsdl2-ttf-dev libboost-all-dev libvorbis-dev libcairo2-dev libpango1.0-dev libssl-dev libreadline-dev cmake make scons pkgconf

Download the source code from https://wiki.wesnoth.org/Download, extract and compile.

tar -xf wesnoth-1.18.0.tar.bz2
cd wesnoth-1.18.0
scons wesnoth

To play the game, you don’t need to install it, simply run in from the same directory:

./wesnoth

Enjoy!

This blog describes how to use Azure libraries (SDK) for Python to list Samba File Shares and read files and folders on Samba File share on Azure.

Overview of Azure SDKs can be found here.

Bicep to create Samba File share

To create SMB File Share using bicep code use Microsoft.Storage storageAccounts/fileServices/shares resource as described here.

The following bicep code creates a file share in an existing storage account:

targetScope = 'resourceGroup'

param storageAccountName string
param fileShareName string


resource storageAccount 'Microsoft.Storage/storageAccounts@2023-01-01' existing = {
  name: storageAccountName
  scope: resourceGroup()
}

resource fileService 'Microsoft.Storage/storageAccounts/fileServices@2021-09-01' = {
  name: 'default'
  parent: storageAccount
}

resource fileShare 'Microsoft.Storage/storageAccounts/fileServices/shares@2023-01-01' = {
  name: fileShareName
  parent: fileService
  properties: {
    enabledProtocols: 'SMB'
  }
}

It can be invoked from a parent bicep file as follows:

param location string = resourceGroup().location
param fileShareName string = 'samba-file-share'
param storageResourceGroupName string = 'rg-storage'
param storageAccountName string = 'sambastorageaccount'

resource storageResourceGroup 'Microsoft.Resources/resourceGroups@2021-04-01' existing = {
  name: storageResourceGroupName
  location: location
}

module fileShares 'file_service_share.bicep' = {
  name: fileShareName
  scope: storageResourceGroup
  params: {
    storageAccountName: storageResourceGroupName
    fileShareName: fileShareName
  }
}

List files and folders in SMB File Share

Authorize requests to Azure Storage

We are going to use role-based access control (RBAC) with Microsoft Entra ID. Azure Storage accepts OAuth 2.0 access tokens from the Microsoft Entra tenant associated with the subscription that contains the storage account. To do so we have to provide appropriate role for the Service Principal.

List of permissions for File service operations can be found here.

Most of the ‘read’ operations require the following roles:

Microsoft.Storage/storageAccounts/fileServices/fileShares/files/read
Microsoft.Storage/storageAccounts/fileServices/readFileBackupSemantics/action

After reviewing list of Azure build in roles it can be found that Storage File Data Privileged Reader role is what we need. It has the following permissions:

"permissions": [
    {
      "actions": [],
      "notActions": [],
      "dataActions": [
        "Microsoft.Storage/storageAccounts/fileServices/fileshares/files/read",
        "Microsoft.Storage/storageAccounts/fileServices/readFileBackupSemantics/action"
      ],
      "notDataActions": []
    }
]

To assign the role to the Service Principal use the following bicep code:

targetScope = 'resourceGroup'

param fileShareName string = 'samba-file-share'
param storageResourceGroupName string = 'rg-storage'
param storageAccountName string = 'sambastorageaccount'
param principalId string = '...' # object id of samba-file-share-sp

var storageFileDataPrivilegedReaderRole = subscriptionResourceId('Microsoft.Authorization/roleDefinitions', 'b8eda974-7b85-4f76-af95-65846b26df6d')

resource storageAccount 'Microsoft.Storage/storageAccounts@2022-09-01' existing = {
  name: storageAccountName
  scope: resourceGroup()
}

resource roleAssignment 'Microsoft.Authorization/roleAssignments@2020-10-01-preview' = {
  name: guid(storageAccountName, principalId, storageFileDataPrivilegedReaderRole)
  scope: storageAccount
  properties: {
    roleDefinitionId: storageFileDataPrivilegedReaderRole
    principalId: principalId
  }
}

Python SDK to access files and folders in File Share

Azure Storage File Share client library for Python used in the code below can be found here.

To run the code install necessary Azure libraries for Python:

pip install azure-identity
pip install azure-storage-file-share

The code below uses ClientSecretCredential to authenticate with Azure AD and ShareClient to access the file share:

from azure.identity import ClientSecretCredential
from azure.storage.fileshare import ShareClient, ShareServiceClient

# Set the storage resource group, account, and share names
storage_account = "sambastorageaccount"
share_name = "samba-file-share"

# Set the Azure AD tenant ID, client ID, and client secret for 'samba-file-share-sp'
tenant_id = "..."
client_id = "..."
client_secret = "..."

# Set the account URL
account_url = f"https://{storage_account}.file.core.windows.net"

# Create a client secret credential
credential = ClientSecretCredential(tenant_id=tenant_id, client_id=client_id, client_secret=client_secret)

# Create a share client
share_client = ShareClient(account_url, share_name, credential=credential, token_intent="backup")

# List directories and files
directories_and_files = list(share_client.list_directories_and_files())
print(f"directories_and_files: {directories_and_files}")

# Get directory properties
directory_client = share_client.get_directory_client("smb_folder")
directory_properties = directory_client.get_directory_properties()
print(f"directory_properties: {directory_properties}")

Sample response is presented below:

directories_and_files: [{'name': 'smb_folder', 'last_modified': None, 'etag': None, 'server_encrypted': None, 'metadata': None, 'change_time': None, 'creation_time': None, 'last_write_time': None, 'last_access_time': None, 'file_attributes': None, 'permission_key': None, 'file_id': '13835128424026341376', 'parent_id': None, 'is_directory': True}]
directory_properties: {'name': None, 'last_modified': datetime.datetime(2024, 3, 8, 15, 12, 7, tzinfo=datetime.timezone.utc), 'etag': '"0x8DC3F821EBF3806"', 'server_encrypted': True, 'metadata': {}, 'change_time': datetime.datetime(2024, 3, 8, 15, 12, 7, 391437), 'creation_time': datetime.datetime(2024, 3, 8, 15, 12, 7, 391437), 'last_write_time': datetime.datetime(2024, 3, 8, 15, 12, 7, 391437), 'last_access_time': None, 'file_attributes': 'Directory', 'permission_key': '13873210695300244643*12131762314841879338', 'file_id': '13835128424026341376', 'parent_id': '0', 'is_directory': True}

Details regarding ShareClient class can be found here.

List SMB File Shares

Authorize requests to Azure Storage

We are going to use role-based access control (RBAC) with Microsoft Entra ID. Azure Storage accepts OAuth 2.0 access tokens from the Microsoft Entra tenant associated with the subscription that contains the storage account. To do so we have to provide appropriate role for the Service Principal.

To list shares we would need the following permissions:

Microsoft.Storage/storageAccounts/fileServices/shares/read

After reviewing list of Azure build in roles it can be found that Reader role is what we need. It has the following permissions:

"permissions": [
    {
      "actions": [
        "*/read"
      ],
      "notActions": [],
      "dataActions": [],
      "notDataActions": []
    }
]

To assign the role to the Service Principal use the following bicep code:

targetScope = 'resourceGroup'

param fileShareName string = 'samba-file-share'
param storageResourceGroupName string = 'rg-storage'
param storageAccountName string = 'sambastorageaccount'
param principalId string = '...' # object id of samba-file-share-sp

var readerRole = subscriptionResourceId('Microsoft.Authorization/roleDefinitions', 'acdd72a7-3385-48ef-bd42-f606fba81ae7')

resource storageAccount 'Microsoft.Storage/storageAccounts@2022-09-01' existing = {
  name: storageAccountName
  scope: resourceGroup()
}

resource roleAssignment 'Microsoft.Authorization/roleAssignments@2020-10-01-preview' = {
  name: guid(storageAccountName, principalId, readerRole)
  scope: storageAccount
  properties: {
    roleDefinitionId: readerRole
    principalId: principalId
  }
}

Python SDK to list File Shares

StorageManagementClient library for Python used in the code below can be found here.

To run the code install necessary Azure libraries for Python:

pip install azure-identity
pip install azure-mgmt-storage

The code below uses ClientSecretCredential to authenticate with Azure AD and StorageManagementClient to list file shares:

from azure.identity import ClientSecretCredential
from azure.mgmt.storage import StorageManagementClient

# Subscription Id
subscription_id = "..."

# Set the storage resource group, account, and share names
storage_resource_group = "rg-storage"
storage_account = "sambastorageaccount"

# Set the Azure AD tenant ID, client ID, and client secret for 'samba-file-share-sp'
tenant_id = "..."
client_id = "..."
client_secret = "..."

# Create a client secret credential
credential = ClientSecretCredential(tenant_id=tenant_id, client_id=client_id, client_secret=client_secret)

# Create a storage management client
client = StorageManagementClient(credential=credential, subscription_id=subscription_id)

# List file shares
response = client.file_shares.list(resource_group_name=storage_resource_group, account_name=storage_account)
for item in response:
    print(item)

Sample response is presented below:

{'additional_properties': {}, 'id': '/subscriptions/.../resourceGroups/rg-storage/providers/Microsoft.Storage/storageAccounts/sambastorageaccount/fileServices/default/shares/samba-file-share', 'name': 'samba-file-share', 'type': 'Microsoft.Storage/storageAccounts/fileServices/shares', 'etag': '"0x8DC4088F239CCBE"', 'last_modified_time': datetime.datetime(2024, 3, 9, 22, 33, 30, tzinfo=<isodate.tzinfo.Utc object at 0x7f4380077490>), 'metadata': None, 'share_quota': 5120, 'enabled_protocols': 'SMB', 'root_squash': None, 'version': None, 'deleted': None, 'deleted_time': None, 'remaining_retention_days': None, 'access_tier': 'TransactionOptimized', 'access_tier_change_time': datetime.datetime(2024, 3, 8, 10, 45, 36, tzinfo=<isodate.tzinfo.Utc object at 0x7f4380077490>), 'access_tier_status': None, 'share_usage_bytes': None, 'lease_status': 'unlocked', 'lease_state': 'available', 'lease_duration': None, 'signed_identifiers': None, 'snapshot_time': None}

Very nice, short (you will do it in 1h) course by DeepLearning.AI & Sharon Zhou - How Diffusion Models Work.

The sample code to train your own model is built using PyTorch. Includes Sharon showing DDIM - denoising diffusion implicit models sampler, much faster than Denoising diffusion probabilistic models (DDPMs).

Co działa:

• Nauka wymaga wysiłku.
• Ćwiczenia w przywoływaniu - fiszki, quizy, testy.
• Refleksja (Co się stało? Co zrobiłem? Jak to rozwiązałem? Co zrobiłbym inaczej następnym razem?).
• Informacja zwrotna dotycząca niewłaściwych odpowiedzi powinna być podana z opóźnieniem.
• Odpowiedzi typu esej lub krótka odpowiedź są lepsze niż testy wielokrotnego wyboru.
• Ćwiczenia rozłożone w czasie. Na tyle długie, aby rozpoczął się proces zapominania, minimum 1 dzień.
• Próba uporania się z problemem, zanim zostaniemy nauczeni jego rozwiązania - generowanie.
• Ćwiczenia przemieszane i zróżnicowane. Zmiana zagadnienia przed jego przećwiczeniem i opanowaniem w pełni.
• Uczenie nowych treści na podstawie już znanych.
• Umieszczanie wiedzy w szerszym kontekście.
• Wydobywanie kluczowych idei z nowego materiału i organizowanie ich.
• Wiara, że zdolności intelektualne można rozwijać. Nastawienie na rozwój.
• Wiara, że popełnianie błędów jest nieodłączną częścią procesu uczenia się.
• Posiadanie wsparcia w rodzicach, wychowawcach, mentorach, którzy w nas wierzą.
• Uczenie się reguł, a nie przykładów.
• Chwalenie za wysiłek, ciężką pracę.
• Odwaga, ciekawość, wytrwałość.
• Mnemotechniki, pałac pamięci.

Co nie działa:

• Ćwiczenia skomasowane, wkuwanie.
• Ćwiczenia zblokowane.
• Wielokrotna lektura tekstu w krótkim czasie to strata czasu.
• Nauka zgodna z preferowanym stylem nauczania (np. słuchowcy vs wzrokowcy).
• Lęk przed niepowodzeniem i popełnianiem błędów. Zaśmiecamy pamięć roboczą na zadręczanie się trudnością otrzymanego zadania.
• Słabo oceniamy, kiedy nauka idzie nam dobrze, a kiedy źle.
• Wiara, że zdolności intelektualne są niezmienne, dziedziczone. Przypisywanie niepowodzeń brakowi zdolności - “nie jestem inteligentny”.
• Koncentracja na osiągnięciu wyniku (zamiast na zdobuciu umiejętności i rozwoju).
• Trudności niepożądane - takie, których nie jesteśmy w stanie pokonać.
• Chwalenie za naturalny talent, np. bystrość.

Proces uczenia się:

• Kodowanie
• Konsolidacja
• Przywoływanie

Przypisy, notatki:

• Specific and varied practice of motor skill - badanie, w którym jedna grupa dzieci ćwiczyła rzuty do kosza z odległości 60 oraz 120cm. Druga grupa ćwiczyła tylko rzuty z odległości 100cm. Po 12 tygodniach obie grupy zostały poddane testowi rzutów z odległości 100cm.
• Encoding difficulty and memory: Toward a unifying theory - utrudnienie czytania tekstu poprzez opuszczanie liter lub nietypową czcionkę opóźnia proces kodowania, ale wpływa pozytywnie na późniejsze przypominanie.
• Intrusion of a thematic idea in retention of prose. Pamięć może ulec zniekształceniu: Helen Keller vs Carol Harris.
• Imagination inflation: Imagining a childhood event inflates confidence that it occurred. Inflacja wyobraźni - “czy kiedykolwiek stłukłeś ręką okno”
• Reconstruction of automobile destruction: An example of the interaction between language and memory. Sugestia - ocena prędkości pojazdów w czasie zderzenia - jeden opis filmu mówił o “kontakie”, drugi o “zderzeniu”.
• Gone but not forgotten: scientists uncover pervasive, unconscious influences on memory. Interferencja - zaatakowana kobieta jako sprawcę podaje mężczyznę, który udzielał wywiadu w telewizji.
• Illusions of Comprehension, Competence, and Remenbering. Przekleństwo wiedzy (curse of knowledge), efekt pewności wstecznej (hindsight bias).
• Social contagion of memory. Dostosowanie (komformizm) pamięci (memory conformity), społeczne skażenie pamięci (social contagion of memory). Efekt fałszywego konensusu (false consensus effect).
• Confessions of a Converted Lecturer: Eric Mazur. Osobą, która najlepiej wie, co sprawia trudność studentowi jest nie profesor, a inny student.
• “Overconfidence in the Communication of Intent: Heard and Unheard Melodies”. Jedna osoba odtwarza w głowie melodię i ją wystukuje, druga zgaduje. Przewidywane poprawne odpowiedzi przez osobę stukającą to 50%, rzeczywiste to 2.5%.
• Policjant w czasie służby odebrał broń napastnikowi, a następnie mu ją oddał.
• 2 zadania: 1. Atak na zamek otoczony fosą i mostami. Mosty są zaminowane, pozwalają przechodzić jedynie małym grupom na raz. Jak zniszczyć guza w mózgu radioterapią, nie uszkadzając zdrowych tkanek.
• Delay of Gratification in Children. Test cukierka (marshmallow study).

Rodzaje inteligencji według Howard Gardner:

• inteligencja przestrzenna
• lingwistyczna
• kinestetyczna
• muzyczna
• interpersonalna
• wewnętrzna
• przyrodnicza

Thanks to Mozilla’s llamafile it is now super easy to run LLM locally.

Install and register ape as per https://github.com/Mozilla-Ocho/llamafile#gotchas.

sudo wget -O /usr/bin/ape https://cosmo.zip/pub/cosmos/bin/ape-$(uname -m).elf
sudo chmod +x /usr/bin/ape
sudo sh -c "echo ':APE:M::MZqFpD::/usr/bin/ape:' >/proc/sys/fs/binfmt_misc/register"
sudo sh -c "echo ':APE-jart:M::jartsr::/usr/bin/ape:' >/proc/sys/fs/binfmt_misc/register"

Download the model from Hugging Face, make it executable and run it:

curl -LO https://huggingface.co/jartine/llava-v1.5-7B-GGUF/resolve/main/llava-v1.5-7b-q4-server.llamafile
chmod +x llava-v1.5-7b-q4-server.llamafile
./llava-v1.5-7b-q4-server.llamafile

Open http://127.0.0.1:8080.

Firma, która nie ma jasnej wizji wykorzystania AI, nie utrzyma się na rynku. Nie jest już potrzebna strategia dla AI, a strategia zawierająca AI (według Strategy For and With AI).

Problem w tym, że opracowania miejsca sztucznej inteligencji w strategii nie da się w pełni zlecić na zewnątrz. Magicy od AI mogą rozhuśtać część technologiczną, acz koniec końców największym wyzwaniem jest to, jaką dźwignię podłożyć pod AI, by zyskać przewagę konkurencyjną w swojej gałęzi rynku i domenie strategicznej (według What’s Your AI Strategy?).

[…] jak podaje Boston Consulting Group firmy, które zyskały dźwignię dzięki skali AI, przeznaczają 10% swojego budżetu AI na algorytmy, 20% na technologie i nawet 70% na upewnienie się, że AI jest dobrze zintegrowana z procesami biznesowymi. Postrzeganie sztucznej inteligencji głównie jako inwestycji w sprzęt czy oprogramowanie to chybiona wizja - najwięcej pracy do wykonania jest w odpowiedniej implementacji.

Końcowa uwaga: nie istnieje uniwersalne podejście do wdrożenia AI w strategii, bo kluczem do sukcesu w jej wdrożeniu jest zdanie sobie sprawy z charakteru i potrzeb firmy.