Dimri

1.2K posts

Dimri banner
Dimri

Dimri

@DimriLogic

Tech consultant, problem-solver, and runner. Sharing my journey and lessons learned. Happy to mentor new talent in IT.

local Katılım Nisan 2023
68 Takip Edilen22 Takipçiler
Roaring Hammy 🏇🌹
Roaring Hammy 🏇🌹@RoaringHammy·
Be honest, when you go to costco, how much do you usually spend? 😵😵
English
19
0
12
1K
Dimri
Dimri@DimriLogic·
@techie_piyush 😀 we need a value in drop down, okay that can be done, n how that new value shd work in flows? silence 🔕,then oh that we still need to figure out lol 😆
English
1
0
1
23
Dimri
Dimri@DimriLogic·
@RealAnkush Have seen other way too, job is posted. However, hire already identified!
English
1
0
0
21
Dimri
Dimri@DimriLogic·
@jamesqquick Mostly these keys are for clients, so shd be more like pk-***** or that been left to clients that don't send sk keys from the front end.
English
0
0
0
3
James Q Quick
James Q Quick@jamesqquick·
I was today years old when I realized that sk in an API key stood for "Secret Key" 🤦‍♂️ For example, in this Anthropic API key "sk-ant-sdk47-24cGL91R43_1f_DSru3ytNg1uN"
English
17
1
73
7.6K
Dimri
Dimri@DimriLogic·
@codemullins N integration test should have unit test assertion to ease debugging for a variance. Limit your test!!
English
0
0
0
50
Cody Mullins
Cody Mullins@codemullins·
you shouldn't write unit tests
English
14
0
12
2.4K
Dr Milan Milanović
Dr Milan Milanović@milan_milanovic·
Software is never done. It evolves, decays, or dies.
English
6
7
60
2.6K
Dimri
Dimri@DimriLogic·
@Narayani07 I thought it would be predictable comfort 😕 and new unpredictable discomfort 😀
English
0
0
1
46
Narayani Gurunathan
Narayani Gurunathan@Narayani07·
You’re not stuck in your career. You’re addicted to predictable discomfort. 😐
English
2
0
33
1.2K
DoItSwift
DoItSwift@DoItSwift_com·
One of the most misunderstood parts of LLMs is that temperature doesn’t add creativity directly - it reshapes the probability distribution before token sampling. Higher temperature flattens probabilities -> more randomness. Lower temperature sharpens them -> more deterministic outputs. Understanding logits and softmax changes how you think about hallucinations and model behavior entirely.
English
1
0
1
125
Arpit Bhayani
Arpit Bhayani@arpit_bhayani·
I always wondered, when we deal with LLMs, how a temperature parameter brings in variance? Yesterday, I got some time to dig into the internals, and all my learnings around it are compiled in my latest blog. By the way, the blog covers this, the basics, and some math in an easy-to-digest manner. In the essay, I broke down how LLMs actually work at the systems level and explained: - what LLMs are mechanically doing - how next-token prediction really works - what logits and softmax are actually doing internally - why identical prompts can produce different outputs - why hallucinations happen - and how temperature really works under the hood by reshaping the probability distribution before sampling If you are building with LLMs (which I am sure you are), understanding these mechanics gives you a much better mental model. Give it a read, and like always, I hope this helps.
Arpit Bhayani tweet media
English
21
71
823
49.3K
Dimri
Dimri@DimriLogic·
@vibrantachintya @arpit_bhayani Almost there + few more tokens so generated text makes sense. If u explore token costing- it's talk about input token + processing tokens
English
0
0
0
10
Achintya अचिन्त्य गौमत
@arpit_bhayani I’m just thinking out loud a very basic aspect. LLMs work by just predicting the next token at a time. So if an output 100 tokens (let’s say), so it means 100 iterations of prompt through whole of LLM layers? Is it correct understanding?
English
1
0
1
138
Dimri
Dimri@DimriLogic·
@Hiteshdotcom Interesting, will be more like one platform for social - text , video etc. Q is how to keep folks interested in reading when, on the other hand, I can listen n view
English
0
0
0
126
Hitesh Choudhary
Hitesh Choudhary@Hiteshdotcom·
X will be a video platform by the end of this year. I really want to be wrong.
English
48
1
313
7.3K
Dimri
Dimri@DimriLogic·
@0fem_15 I think mainly due to transmission between router(very close) vs. mobile tower 🤔 , the far source needs strong signals n continuous data packets
English
0
0
2
596
𝐎𝐟𝐞𝐦
𝐎𝐟𝐞𝐦@0fem_15·
The difference in battery drain between using data and Wi-Fi is actually crazy.
English
294
2.5K
45K
2M
Dimri
Dimri@DimriLogic·
@Gubloinvestor Middle/upper class! Wow, no rent or mortgage ,congratulations to the couple 😀, otherwise different story
English
2
0
1
137
Gublo 🇨🇦
Gublo 🇨🇦@Gubloinvestor·
How would you rate this lifestyle expenses for couple and 2 infants In Nova scotia, Canada. Groceries $1000 Baby supplies $1000 Car insurance $400 Gas $400 Property taxes $400 Entertainment $300 Electricity $300 Propane $250 Phone bills $150 Home insurance $100 Lawn& Snow $100 Wifi $100 Others $500 Total $5000 a month Lower class, middle class, upper middle class or anything else?
English
69
4
58
21.3K
CG Investing
CG Investing@CGInvesting10·
For the non $SOFI holders… At what price would you consider buying $SOFI?
CG Investing tweet media
English
177
23
305
137.1K
Dimri
Dimri@DimriLogic·
@gurgavin N we solve geo location problems, during sunlight your speed will be amazing n degrade after sunset, CDN are at risk lol
English
1
0
1
133
GURGAVIN
GURGAVIN@gurgavin·
JUST IN : GOOGLE & SPACEX ARE IN TALKS TO LAUNCH DATA CENTERS IN ORBIT PER WSJ $GOOGL
English
53
70
1.6K
125.8K
Dimri
Dimri@DimriLogic·
@LegacyPathCA It's also like a good habit. Once u start, u like it 😀 I started more like experiment n happy with it
English
0
0
0
113
Erik | Legacy Path 🇨🇦
Two Canadians. Canadian A opened a TFSA at 18. Put $200/month into a global ETF. Never thought about it again. Canadian B waited until 30 to "figure out investing first." At 65: Canadian A → $1,250,000. Tax free. Canadian B → $450,000. Tax free. The best time to open it was at 18. The second best time is today.
English
37
21
380
78.6K
Dimri
Dimri@DimriLogic·
@ShawnDevDedalus "Still with us" the key findings there 😀, fresh eyes 👀 catch more bugs lol
English
0
0
1
6
Shawn Dedalus .·. 🇺🇲 ⚓
We found a bug from 2 years ago, fixed it, and now 3 new ones appear. This is what I call bug reproduction magic. During the meeting with my devs, we laughed a lot about this. The positive thing is that all the people who worked on this project are still with us. 😎
English
6
1
13
324
Dimri
Dimri@DimriLogic·
@pizdets17 Innovation champs 🏆, no other tech platform, can match them. I m yet to move all my stuff 🙃, rrsp and tfsa is with them 😀, n will get credit card 💳
English
0
0
0
246
WarrantWhisperer
WarrantWhisperer@pizdets17·
thinking of transferring all my TD stuff (rrsp, tfsa, direct investing, checking account, credit card) to Wealthsimple. Thoughts?
English
335
15
1.1K
334.5K
Jon Flynn | Real Estate Data & Analysis
Looking for a likeminded agent who services Richmond Hill & Markham to work with some first time buyers (townhouse). Hit me up if you know someone honest who’s knows numbers & will work in their best interest & won’t sell them out for a quick sale.
English
16
1
43
7.4K
Dimri
Dimri@DimriLogic·
@RealAnkush we should put an error message more cryptic - our agent been notified , agent can be human or AI agent lol 😆
English
0
0
1
97
Ankush Dharkar wants you on tier3.college
Great. Now even google is having reliability issues How's AI coding treating you? Guess software for real world is much more than just code generation afterall!
Ankush Dharkar wants you on tier3.college tweet media
English
12
0
90
5.6K